diff --git a/.DS_Store b/.DS_Store new file mode 100644 index 0000000..2d0118d Binary files /dev/null and b/.DS_Store differ diff --git a/LICENSE b/LICENSE new file mode 100755 index 0000000..4f5d0f1 --- /dev/null +++ b/LICENSE @@ -0,0 +1,673 @@ + GNU GENERAL PUBLIC LICENSE + Version 3, 29 June 2007 + + Copyright (C) 2007 Free Software Foundation, Inc. + Everyone is permitted to copy and distribute verbatim copies + of this license document, but changing it is not allowed. + + Preamble + + The GNU General Public License is a free, copyleft license for +software and other kinds of works. + + The licenses for most software and other practical works are designed +to take away your freedom to share and change the works. By contrast, +the GNU General Public License is intended to guarantee your freedom to +share and change all versions of a program--to make sure it remains free +software for all its users. We, the Free Software Foundation, use the +GNU General Public License for most of our software; it applies also to +any other work released this way by its authors. You can apply it to +your programs, too. + + When we speak of free software, we are referring to freedom, not +price. Our General Public Licenses are designed to make sure that you +have the freedom to distribute copies of free software (and charge for +them if you wish), that you receive source code or can get it if you +want it, that you can change the software or use pieces of it in new +free programs, and that you know you can do these things. + + To protect your rights, we need to prevent others from denying you +these rights or asking you to surrender the rights. Therefore, you have +certain responsibilities if you distribute copies of the software, or if +you modify it: responsibilities to respect the freedom of others. + + For example, if you distribute copies of such a program, whether +gratis or for a fee, you must pass on to the recipients the same +freedoms that you received. You must make sure that they, too, receive +or can get the source code. And you must show them these terms so they +know their rights. + + Developers that use the GNU GPL protect your rights with two steps: +(1) assert copyright on the software, and (2) offer you this License +giving you legal permission to copy, distribute and/or modify it. + + For the developers' and authors' protection, the GPL clearly explains +that there is no warranty for this free software. For both users' and +authors' sake, the GPL requires that modified versions be marked as +changed, so that their problems will not be attributed erroneously to +authors of previous versions. + + Some devices are designed to deny users access to install or run +modified versions of the software inside them, although the manufacturer +can do so. This is fundamentally incompatible with the aim of +protecting users' freedom to change the software. The systematic +pattern of such abuse occurs in the area of products for individuals to +use, which is precisely where it is most unacceptable. Therefore, we +have designed this version of the GPL to prohibit the practice for those +products. If such problems arise substantially in other domains, we +stand ready to extend this provision to those domains in future versions +of the GPL, as needed to protect the freedom of users. + + Finally, every program is threatened constantly by software patents. +States should not allow patents to restrict development and use of +software on general-purpose computers, but in those that do, we wish to +avoid the special danger that patents applied to a free program could +make it effectively proprietary. To prevent this, the GPL assures that +patents cannot be used to render the program non-free. + + The precise terms and conditions for copying, distribution and +modification follow. + + TERMS AND CONDITIONS + + 0. Definitions. + + "This License" refers to version 3 of the GNU General Public License. + + "Copyright" also means copyright-like laws that apply to other kinds of +works, such as semiconductor masks. + + "The Program" refers to any copyrightable work licensed under this +License. Each licensee is addressed as "you". "Licensees" and +"recipients" may be individuals or organizations. + + To "modify" a work means to copy from or adapt all or part of the work +in a fashion requiring copyright permission, other than the making of an +exact copy. The resulting work is called a "modified version" of the +earlier work or a work "based on" the earlier work. + + A "covered work" means either the unmodified Program or a work based +on the Program. + + To "propagate" a work means to do anything with it that, without +permission, would make you directly or secondarily liable for +infringement under applicable copyright law, except executing it on a +computer or modifying a private copy. Propagation includes copying, +distribution (with or without modification), making available to the +public, and in some countries other activities as well. + + To "convey" a work means any kind of propagation that enables other +parties to make or receive copies. Mere interaction with a user through +a computer network, with no transfer of a copy, is not conveying. + + An interactive user interface displays "Appropriate Legal Notices" +to the extent that it includes a convenient and prominently visible +feature that (1) displays an appropriate copyright notice, and (2) +tells the user that there is no warranty for the work (except to the +extent that warranties are provided), that licensees may convey the +work under this License, and how to view a copy of this License. If +the interface presents a list of user commands or options, such as a +menu, a prominent item in the list meets this criterion. + + 1. Source Code. + + The "source code" for a work means the preferred form of the work +for making modifications to it. "Object code" means any non-source +form of a work. + + A "Standard Interface" means an interface that either is an official +standard defined by a recognized standards body, or, in the case of +interfaces specified for a particular programming language, one that +is widely used among developers working in that language. + + The "System Libraries" of an executable work include anything, other +than the work as a whole, that (a) is included in the normal form of +packaging a Major Component, but which is not part of that Major +Component, and (b) serves only to enable use of the work with that +Major Component, or to implement a Standard Interface for which an +implementation is available to the public in source code form. A +"Major Component", in this context, means a major essential component +(kernel, window system, and so on) of the specific operating system +(if any) on which the executable work runs, or a compiler used to +produce the work, or an object code interpreter used to run it. + + The "Corresponding Source" for a work in object code form means all +the source code needed to generate, install, and (for an executable +work) run the object code and to modify the work, including scripts to +control those activities. However, it does not include the work's +System Libraries, or general-purpose tools or generally available free +programs which are used unmodified in performing those activities but +which are not part of the work. For example, Corresponding Source +includes interface definition files associated with source files for +the work, and the source code for shared libraries and dynamically +linked subprograms that the work is specifically designed to require, +such as by intimate data communication or control flow between those +subprograms and other parts of the work. + + The Corresponding Source need not include anything that users +can regenerate automatically from other parts of the Corresponding +Source. + + The Corresponding Source for a work in source code form is that +same work. + + 2. Basic Permissions. + + All rights granted under this License are granted for the term of +copyright on the Program, and are irrevocable provided the stated +conditions are met. This License explicitly affirms your unlimited +permission to run the unmodified Program. The output from running a +covered work is covered by this License only if the output, given its +content, constitutes a covered work. This License acknowledges your +rights of fair use or other equivalent, as provided by copyright law. + + You may make, run and propagate covered works that you do not +convey, without conditions so long as your license otherwise remains +in force. You may convey covered works to others for the sole purpose +of having them make modifications exclusively for you, or provide you +with facilities for running those works, provided that you comply with +the terms of this License in conveying all material for which you do +not control copyright. Those thus making or running the covered works +for you must do so exclusively on your behalf, under your direction +and control, on terms that prohibit them from making any copies of +your copyrighted material outside their relationship with you. + + Conveying under any other circumstances is permitted solely under +the conditions stated below. Sublicensing is not allowed; section 10 +makes it unnecessary. + + 3. Protecting Users' Legal Rights From Anti-Circumvention Law. + + No covered work shall be deemed part of an effective technological +measure under any applicable law fulfilling obligations under article +11 of the WIPO copyright treaty adopted on 20 December 1996, or +similar laws prohibiting or restricting circumvention of such +measures. + + When you convey a covered work, you waive any legal power to forbid +circumvention of technological measures to the extent such circumvention +is effected by exercising rights under this License with respect to +the covered work, and you disclaim any intention to limit operation or +modification of the work as a means of enforcing, against the work's +users, your or third parties' legal rights to forbid circumvention of +technological measures. + + 4. Conveying Verbatim Copies. + + You may convey verbatim copies of the Program's source code as you +receive it, in any medium, provided that you conspicuously and +appropriately publish on each copy an appropriate copyright notice; +keep intact all notices stating that this License and any +non-permissive terms added in accord with section 7 apply to the code; +keep intact all notices of the absence of any warranty; and give all +recipients a copy of this License along with the Program. + + You may charge any price or no price for each copy that you convey, +and you may offer support or warranty protection for a fee. + + 5. Conveying Modified Source Versions. + + You may convey a work based on the Program, or the modifications to +produce it from the Program, in the form of source code under the +terms of section 4, provided that you also meet all of these conditions: + + a) The work must carry prominent notices stating that you modified + it, and giving a relevant date. + + b) The work must carry prominent notices stating that it is + released under this License and any conditions added under section + 7. This requirement modifies the requirement in section 4 to + "keep intact all notices". + + c) You must license the entire work, as a whole, under this + License to anyone who comes into possession of a copy. This + License will therefore apply, along with any applicable section 7 + additional terms, to the whole of the work, and all its parts, + regardless of how they are packaged. This License gives no + permission to license the work in any other way, but it does not + invalidate such permission if you have separately received it. + + d) If the work has interactive user interfaces, each must display + Appropriate Legal Notices; however, if the Program has interactive + interfaces that do not display Appropriate Legal Notices, your + work need not make them do so. + + A compilation of a covered work with other separate and independent +works, which are not by their nature extensions of the covered work, +and which are not combined with it such as to form a larger program, +in or on a volume of a storage or distribution medium, is called an +"aggregate" if the compilation and its resulting copyright are not +used to limit the access or legal rights of the compilation's users +beyond what the individual works permit. Inclusion of a covered work +in an aggregate does not cause this License to apply to the other +parts of the aggregate. + + 6. Conveying Non-Source Forms. + + You may convey a covered work in object code form under the terms +of sections 4 and 5, provided that you also convey the +machine-readable Corresponding Source under the terms of this License, +in one of these ways: + + a) Convey the object code in, or embodied in, a physical product + (including a physical distribution medium), accompanied by the + Corresponding Source fixed on a durable physical medium + customarily used for software interchange. + + b) Convey the object code in, or embodied in, a physical product + (including a physical distribution medium), accompanied by a + written offer, valid for at least three years and valid for as + long as you offer spare parts or customer support for that product + model, to give anyone who possesses the object code either (1) a + copy of the Corresponding Source for all the software in the + product that is covered by this License, on a durable physical + medium customarily used for software interchange, for a price no + more than your reasonable cost of physically performing this + conveying of source, or (2) access to copy the + Corresponding Source from a network server at no charge. + + c) Convey individual copies of the object code with a copy of the + written offer to provide the Corresponding Source. This + alternative is allowed only occasionally and noncommercially, and + only if you received the object code with such an offer, in accord + with subsection 6b. + + d) Convey the object code by offering access from a designated + place (gratis or for a charge), and offer equivalent access to the + Corresponding Source in the same way through the same place at no + further charge. You need not require recipients to copy the + Corresponding Source along with the object code. If the place to + copy the object code is a network server, the Corresponding Source + may be on a different server (operated by you or a third party) + that supports equivalent copying facilities, provided you maintain + clear directions next to the object code saying where to find the + Corresponding Source. Regardless of what server hosts the + Corresponding Source, you remain obligated to ensure that it is + available for as long as needed to satisfy these requirements. + + e) Convey the object code using peer-to-peer transmission, provided + you inform other peers where the object code and Corresponding + Source of the work are being offered to the general public at no + charge under subsection 6d. + + A separable portion of the object code, whose source code is excluded +from the Corresponding Source as a System Library, need not be +included in conveying the object code work. + + A "User Product" is either (1) a "consumer product", which means any +tangible personal property which is normally used for personal, family, +or household purposes, or (2) anything designed or sold for incorporation +into a dwelling. In determining whether a product is a consumer product, +doubtful cases shall be resolved in favor of coverage. For a particular +product received by a particular user, "normally used" refers to a +typical or common use of that class of product, regardless of the status +of the particular user or of the way in which the particular user +actually uses, or expects or is expected to use, the product. A product +is a consumer product regardless of whether the product has substantial +commercial, industrial or non-consumer uses, unless such uses represent +the only significant mode of use of the product. + + "Installation Information" for a User Product means any methods, +procedures, authorization keys, or other information required to install +and execute modified versions of a covered work in that User Product from +a modified version of its Corresponding Source. The information must +suffice to ensure that the continued functioning of the modified object +code is in no case prevented or interfered with solely because +modification has been made. + + If you convey an object code work under this section in, or with, or +specifically for use in, a User Product, and the conveying occurs as +part of a transaction in which the right of possession and use of the +User Product is transferred to the recipient in perpetuity or for a +fixed term (regardless of how the transaction is characterized), the +Corresponding Source conveyed under this section must be accompanied +by the Installation Information. But this requirement does not apply +if neither you nor any third party retains the ability to install +modified object code on the User Product (for example, the work has +been installed in ROM). + + The requirement to provide Installation Information does not include a +requirement to continue to provide support service, warranty, or updates +for a work that has been modified or installed by the recipient, or for +the User Product in which it has been modified or installed. Access to a +network may be denied when the modification itself materially and +adversely affects the operation of the network or violates the rules and +protocols for communication across the network. + + Corresponding Source conveyed, and Installation Information provided, +in accord with this section must be in a format that is publicly +documented (and with an implementation available to the public in +source code form), and must require no special password or key for +unpacking, reading or copying. + + 7. Additional Terms. + + "Additional permissions" are terms that supplement the terms of this +License by making exceptions from one or more of its conditions. +Additional permissions that are applicable to the entire Program shall +be treated as though they were included in this License, to the extent +that they are valid under applicable law. If additional permissions +apply only to part of the Program, that part may be used separately +under those permissions, but the entire Program remains governed by +this License without regard to the additional permissions. + + When you convey a copy of a covered work, you may at your option +remove any additional permissions from that copy, or from any part of +it. (Additional permissions may be written to require their own +removal in certain cases when you modify the work.) You may place +additional permissions on material, added by you to a covered work, +for which you have or can give appropriate copyright permission. + + Notwithstanding any other provision of this License, for material you +add to a covered work, you may (if authorized by the copyright holders of +that material) supplement the terms of this License with terms: + + a) Disclaiming warranty or limiting liability differently from the + terms of sections 15 and 16 of this License; or + + b) Requiring preservation of specified reasonable legal notices or + author attributions in that material or in the Appropriate Legal + Notices displayed by works containing it; or + + c) Prohibiting misrepresentation of the origin of that material, or + requiring that modified versions of such material be marked in + reasonable ways as different from the original version; or + + d) Limiting the use for publicity purposes of names of licensors or + authors of the material; or + + e) Declining to grant rights under trademark law for use of some + trade names, trademarks, or service marks; or + + f) Requiring indemnification of licensors and authors of that + material by anyone who conveys the material (or modified versions of + it) with contractual assumptions of liability to the recipient, for + any liability that these contractual assumptions directly impose on + those licensors and authors. + + All other non-permissive additional terms are considered "further +restrictions" within the meaning of section 10. If the Program as you +received it, or any part of it, contains a notice stating that it is +governed by this License along with a term that is a further +restriction, you may remove that term. If a license document contains +a further restriction but permits relicensing or conveying under this +License, you may add to a covered work material governed by the terms +of that license document, provided that the further restriction does +not survive such relicensing or conveying. + + If you add terms to a covered work in accord with this section, you +must place, in the relevant source files, a statement of the +additional terms that apply to those files, or a notice indicating +where to find the applicable terms. + + Additional terms, permissive or non-permissive, may be stated in the +form of a separately written license, or stated as exceptions; +the above requirements apply either way. + + 8. Termination. + + You may not propagate or modify a covered work except as expressly +provided under this License. Any attempt otherwise to propagate or +modify it is void, and will automatically terminate your rights under +this License (including any patent licenses granted under the third +paragraph of section 11). + + However, if you cease all violation of this License, then your +license from a particular copyright holder is reinstated (a) +provisionally, unless and until the copyright holder explicitly and +finally terminates your license, and (b) permanently, if the copyright +holder fails to notify you of the violation by some reasonable means +prior to 60 days after the cessation. + + Moreover, your license from a particular copyright holder is +reinstated permanently if the copyright holder notifies you of the +violation by some reasonable means, this is the first time you have +received notice of violation of this License (for any work) from that +copyright holder, and you cure the violation prior to 30 days after +your receipt of the notice. + + Termination of your rights under this section does not terminate the +licenses of parties who have received copies or rights from you under +this License. If your rights have been terminated and not permanently +reinstated, you do not qualify to receive new licenses for the same +material under section 10. + + 9. Acceptance Not Required for Having Copies. + + You are not required to accept this License in order to receive or +run a copy of the Program. Ancillary propagation of a covered work +occurring solely as a consequence of using peer-to-peer transmission +to receive a copy likewise does not require acceptance. However, +nothing other than this License grants you permission to propagate or +modify any covered work. These actions infringe copyright if you do +not accept this License. Therefore, by modifying or propagating a +covered work, you indicate your acceptance of this License to do so. + + 10. Automatic Licensing of Downstream Recipients. + + Each time you convey a covered work, the recipient automatically +receives a license from the original licensors, to run, modify and +propagate that work, subject to this License. You are not responsible +for enforcing compliance by third parties with this License. + + An "entity transaction" is a transaction transferring control of an +organization, or substantially all assets of one, or subdividing an +organization, or merging organizations. If propagation of a covered +work results from an entity transaction, each party to that +transaction who receives a copy of the work also receives whatever +licenses to the work the party's predecessor in interest had or could +give under the previous paragraph, plus a right to possession of the +Corresponding Source of the work from the predecessor in interest, if +the predecessor has it or can get it with reasonable efforts. + + You may not impose any further restrictions on the exercise of the +rights granted or affirmed under this License. For example, you may +not impose a license fee, royalty, or other charge for exercise of +rights granted under this License, and you may not initiate litigation +(including a cross-claim or counterclaim in a lawsuit) alleging that +any patent claim is infringed by making, using, selling, offering for +sale, or importing the Program or any portion of it. + + 11. Patents. + + A "contributor" is a copyright holder who authorizes use under this +License of the Program or a work on which the Program is based. The +work thus licensed is called the contributor's "contributor version". + + A contributor's "essential patent claims" are all patent claims +owned or controlled by the contributor, whether already acquired or +hereafter acquired, that would be infringed by some manner, permitted +by this License, of making, using, or selling its contributor version, +but do not include claims that would be infringed only as a +consequence of further modification of the contributor version. For +purposes of this definition, "control" includes the right to grant +patent sublicenses in a manner consistent with the requirements of +this License. + + Each contributor grants you a non-exclusive, worldwide, royalty-free +patent license under the contributor's essential patent claims, to +make, use, sell, offer for sale, import and otherwise run, modify and +propagate the contents of its contributor version. + + In the following three paragraphs, a "patent license" is any express +agreement or commitment, however denominated, not to enforce a patent +(such as an express permission to practice a patent or covenant not to +sue for patent infringement). To "grant" such a patent license to a +party means to make such an agreement or commitment not to enforce a +patent against the party. + + If you convey a covered work, knowingly relying on a patent license, +and the Corresponding Source of the work is not available for anyone +to copy, free of charge and under the terms of this License, through a +publicly available network server or other readily accessible means, +then you must either (1) cause the Corresponding Source to be so +available, or (2) arrange to deprive yourself of the benefit of the +patent license for this particular work, or (3) arrange, in a manner +consistent with the requirements of this License, to extend the patent +license to downstream recipients. "Knowingly relying" means you have +actual knowledge that, but for the patent license, your conveying the +covered work in a country, or your recipient's use of the covered work +in a country, would infringe one or more identifiable patents in that +country that you have reason to believe are valid. + + If, pursuant to or in connection with a single transaction or +arrangement, you convey, or propagate by procuring conveyance of, a +covered work, and grant a patent license to some of the parties +receiving the covered work authorizing them to use, propagate, modify +or convey a specific copy of the covered work, then the patent license +you grant is automatically extended to all recipients of the covered +work and works based on it. + + A patent license is "discriminatory" if it does not include within +the scope of its coverage, prohibits the exercise of, or is +conditioned on the non-exercise of one or more of the rights that are +specifically granted under this License. You may not convey a covered +work if you are a party to an arrangement with a third party that is +in the business of distributing software, under which you make payment +to the third party based on the extent of your activity of conveying +the work, and under which the third party grants, to any of the +parties who would receive the covered work from you, a discriminatory +patent license (a) in connection with copies of the covered work +conveyed by you (or copies made from those copies), or (b) primarily +for and in connection with specific products or compilations that +contain the covered work, unless you entered into that arrangement, +or that patent license was granted, prior to 28 March 2007. + + Nothing in this License shall be construed as excluding or limiting +any implied license or other defenses to infringement that may +otherwise be available to you under applicable patent law. + + 12. No Surrender of Others' Freedom. + + If conditions are imposed on you (whether by court order, agreement or +otherwise) that contradict the conditions of this License, they do not +excuse you from the conditions of this License. If you cannot convey a +covered work so as to satisfy simultaneously your obligations under this +License and any other pertinent obligations, then as a consequence you may +not convey it at all. For example, if you agree to terms that obligate you +to collect a royalty for further conveying from those to whom you convey +the Program, the only way you could satisfy both those terms and this +License would be to refrain entirely from conveying the Program. + + 13. Use with the GNU Affero General Public License. + + Notwithstanding any other provision of this License, you have +permission to link or combine any covered work with a work licensed +under version 3 of the GNU Affero General Public License into a single +combined work, and to convey the resulting work. The terms of this +License will continue to apply to the part which is the covered work, +but the special requirements of the GNU Affero General Public License, +section 13, concerning interaction through a network will apply to the +combination as such. + + 14. Revised Versions of this License. + + The Free Software Foundation may publish revised and/or new versions of +the GNU General Public License from time to time. Such new versions will +be similar in spirit to the present version, but may differ in detail to +address new problems or concerns. + + Each version is given a distinguishing version number. If the +Program specifies that a certain numbered version of the GNU General +Public License "or any later version" applies to it, you have the +option of following the terms and conditions either of that numbered +version or of any later version published by the Free Software +Foundation. If the Program does not specify a version number of the +GNU General Public License, you may choose any version ever published +by the Free Software Foundation. + + If the Program specifies that a proxy can decide which future +versions of the GNU General Public License can be used, that proxy's +public statement of acceptance of a version permanently authorizes you +to choose that version for the Program. + + Later license versions may give you additional or different +permissions. However, no additional obligations are imposed on any +author or copyright holder as a result of your choosing to follow a +later version. + + 15. Disclaimer of Warranty. + + THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY +APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT +HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY +OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, +THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR +PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM +IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF +ALL NECESSARY SERVICING, REPAIR OR CORRECTION. + + 16. Limitation of Liability. + + IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS +THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY +GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE +USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF +DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD +PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), +EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF +SUCH DAMAGES. + + 17. Interpretation of Sections 15 and 16. + + If the disclaimer of warranty and limitation of liability provided +above cannot be given local legal effect according to their terms, +reviewing courts shall apply local law that most closely approximates +an absolute waiver of all civil liability in connection with the +Program, unless a warranty or assumption of liability accompanies a +copy of the Program in return for a fee. + + END OF TERMS AND CONDITIONS + + How to Apply These Terms to Your New Programs + + If you develop a new program, and you want it to be of the greatest +possible use to the public, the best way to achieve this is to make it +free software which everyone can redistribute and change under these terms. + + To do so, attach the following notices to the program. It is safest +to attach them to the start of each source file to most effectively +state the exclusion of warranty; and each file should have at least +the "copyright" line and a pointer to where the full notice is found. + + Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . + +Also add information on how to contact you by electronic and paper mail. + + If the program does terminal interaction, make it output a short +notice like this when it starts in an interactive mode: + + Copyright (C) + This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. + This is free software, and you are welcome to redistribute it + under certain conditions; type `show c' for details. + +The hypothetical commands `show w' and `show c' should show the appropriate +parts of the General Public License. Of course, your program's commands +might be different; for a GUI interface, you would use an "about box". + + You should also get your employer (if you work as a programmer) or school, +if any, to sign a "copyright disclaimer" for the program, if necessary. +For more information on this, and how to apply and follow the GNU GPL, see +. + + The GNU General Public License does not permit incorporating your program +into proprietary programs. If your program is a subroutine library, you +may consider it more useful to permit linking proprietary applications with +the library. If this is what you want to do, use the GNU Lesser General +Public License instead of this License. But first, please read +. diff --git a/LICENSES_THIRD_PARTY b/LICENSES_THIRD_PARTY new file mode 100644 index 0000000..3917cb3 --- /dev/null +++ b/LICENSES_THIRD_PARTY @@ -0,0 +1,20 @@ +-------------------------------------------------- +Third party dependencies listed by License type +[Format: name (Python module) - URL] +-------------------------------------------------- + +OSI Approved (new BSD) +* SciPy (scipy) - https://github.com/scipy/scipy/blob/main/LICENSE.txt +* scikit-learn (sklearn) - https://github.com/scikit-learn/scikit-learn/blob/main/COPYING + +BSD 3-Clause License +* Numpy (numpy) - https://github.com/numpy/numpy/blob/main/LICENSE.txt +* Pandas (pandas) - https://github.com/pandas-dev/pandas/blob/5aba6659e422e985683cfb46c07c3364a02b6e5b/AUTHORS.md +* Dill (dill) - https://github.com/uqfoundation/dill/blob/master/LICENSE + +MIT License (MIT) +* Keras (keras) - https://github.com/keras-team/keras/blob/dc698c5486117780b643eda0a2f60a8753625b8a/LICENSE +* LightGBM (lightgbm) - https://github.com/microsoft/LightGBM/blob/master/LICENSE + +Apache Software License (Apache 2.0) +* TensorFlow (tensorflow) - https://github.com/tensorflow/tensorflow/blob/6b6d843ccab78f9f91c3b98a43ca09ffecad4747/LICENSE diff --git a/README.md b/README.md new file mode 100755 index 0000000..e0a0a7d --- /dev/null +++ b/README.md @@ -0,0 +1,44 @@ +# Prediction Uncertainty for QSAR + +This package contains Python code to construct prediction intervals for QSAR regression. +The implemented QSAR prediction models include: Random Forests, Fully-Connected Neural Networks, and Gradient Boosting. +The methodology for developing prediction intervals accompanying the point predictors can be find in Reference [1] and [2]. + +Authors: Yuting Xu, Andy Liaw, Robert P. Sheridan, and Vladimir Svetnik + +Affiliation: Merck & Co., Inc., Rahway, New Jersey 07065, United States + +Maintainer: yuting.xu@merck.com + +Last updated: 08/29/2023 + +## Workflow + +

+ Logo_workflow +

+ +## Usage + +The code is written in the functional programming paradigm without the hassle of installation. +Simply clone or download the repository to your local machine, and use the provided examples as a starting point to experiment with your own workflow. + +### Prerequisites + +* numpy +* pandas +* dill +* scipy +* scikit-learn +* tensorflow +* keras +* lightgbm + +## Reference + +[1] Xu, Y., Liaw, A., Sheridan, R.P. and Svetnik, V., 2023. Development and Evaluation of Conformal Prediction Methods for QSAR. arXiv preprint arXiv:2304.00970. + +[2] Cortes-Ciriano, I.; Bender, A. Reliable prediction errors for deep neural networks using test-time dropout. Journal of chemical information and modeling 2019, 59, 3330–3339. + +## License +This project is licensed under the GNU General Public License v3.0 License - see the [LICENSE](LICENSE) file for details. diff --git a/data/README.md b/data/README.md new file mode 100755 index 0000000..50cb407 --- /dev/null +++ b/data/README.md @@ -0,0 +1,3 @@ +## Unzip the compressed example datasets in this folder + +tar -xvf exampleData.tar.xz diff --git a/data/exampleData.tar.xz b/data/exampleData.tar.xz new file mode 100755 index 0000000..ff8ae74 Binary files /dev/null and b/data/exampleData.tar.xz differ diff --git a/docs/readme_logo.jpg b/docs/readme_logo.jpg new file mode 100755 index 0000000..d2ffe16 Binary files /dev/null and b/docs/readme_logo.jpg differ diff --git a/examples/run_DNN-dropout.py b/examples/run_DNN-dropout.py new file mode 100755 index 0000000..7450d28 --- /dev/null +++ b/examples/run_DNN-dropout.py @@ -0,0 +1,107 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +#!/usr/bin/env python + +# CUDA_VISIBLE_DEVICES=7 python run_DNN_dropout.py + +# Prerequisites +import os +import sys +import time +import glob +import numpy as np +import pandas as pd +import dill +sys.path.append("../puqsar") + +# Create folder to save results +saveFolder = '../results/DNN-dropout' +if not os.path.exists(saveFolder): + os.makedirs(saveFolder) + +# Load Example Data +dat_train = pd.read_csv("../data/3A4_processed/dat_train.csv") +dat_test= pd.read_csv("../data/3A4_processed/dat_test_rs.csv") + +# Preparing Data for DNN-dropout model training and testing +from utils.preprocessing import * +X_train, X_cal, y_train_norm, y_cal_norm, mu_tr, sd_tr, df_label_train, df_label_cal = preprocessing_DNN_default_train(dat_train, p_cal = 0.2, seed=111) +X_test, df_label_test = preprocessing_DNN_default_test(dat_test, mu_tr, sd_tr) + +# Load functions for DNN dropout model +from models.DNN_dropout import * + +# Hyperparameters for DNN structure and training +p_batchSize = 0.05 +learn_rate = 0.001 + +nn_pars = {'nodes' : [4000, 2000, 1000, 1000], + 'dropout': [0.25, 0.25, 0.25, 0.1], + 'batch_size' : min(128, round(X_train.shape[0]*p_batchSize)), + 'learn_rate' : learn_rate, + 'epochs' : 500, + 'wt_decay' : 0.00005} +dropouts = 100 + +# Train a DNN-dropout model +model = train_DNN_dropout(X_train, y_train_norm, X_cal, y_cal_norm, nn_pars) + +# Prediction on Calibration set +pred_cal = np.zeros((X_cal.shape[0],dropouts)) +for k in range(dropouts): + pred_cal[:,k] = model.predict(X_cal)[:, 0] + +pred_cal_avg = np.mean(pred_cal, 1) * sd_tr + mu_tr +pred_cal_sd = np.std(pred_cal, 1) * sd_tr +df_pred_cal = pd.concat([df_label_cal.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_cal_avg, "Pred_UNC": pred_cal_sd})], axis=1, ignore_index = False, sort = False) + +# Calibration step +# Conformal algorithm options: CP_ACE, CP_expSD, CP_homo +from calibrators.ICP import * +nominal_level=0.8 +fun_PI = CP_ACE(df_pred_cal, nominal_level) + +# Save the model to .h5, prediction for calibration set (including raw unertainty score) to .csv, and the calibration results to .pkl file +model_path = os.path.join(saveFolder, 'model.h5') +model.save(model_path) +df_pred_cal.to_csv(os.path.join(saveFolder,"df_pred_cal.csv"), header=True, index=False) +with open(os.path.join(saveFolder, 'calibration.pkl'), 'wb') as file: + dill.dump([fun_PI,nominal_level,mu_tr,sd_tr], file) + +# Application on Test set and save results as CSV +""" +from tensorflow.keras.models import load_model +model = load_model(os.path.join(saveFolder, 'model.h5')) + +with open(os.path.join(saveFolder, 'calibration.pkl'), 'rb') as file: + fun_PI,nominal_level,mu_tr,sd_tr = dill.load(file) +""" + +pred_test = np.zeros((X_test.shape[0],dropouts)) +for k in range(dropouts): + pred_test[:,k] = model.predict(X_test)[:, 0] + +pred_test_avg = np.mean(pred_test, 1) * sd_tr + mu_tr +pred_test_sd = np.std(pred_test, 1) * sd_tr + +df_pred_test = pd.concat([df_label_test.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_test_avg,"Pred_UNC": pred_test_sd})], axis=1, ignore_index = False, sort = False) +df_pred_test = fun_PI(df_pred_test) +df_pred_test.to_csv(os.path.join(saveFolder,"df_pred_test.csv"), header=True, index=False) diff --git a/examples/run_DNN-multitask.py b/examples/run_DNN-multitask.py new file mode 100755 index 0000000..0609281 --- /dev/null +++ b/examples/run_DNN-multitask.py @@ -0,0 +1,105 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +#!/usr/bin/env python + +# CUDA_VISIBLE_DEVICES=7 python run_DNN_multitask.py + +# Prerequisites +import os +import sys +import time +import glob +import numpy as np +import pandas as pd +import dill +sys.path.append("../puqsar") + +# Create folder to save results +saveFolder = '../results/DNN-multitask' +if not os.path.exists(saveFolder): + os.makedirs(saveFolder) + +# Load Example Data +dat_train = pd.read_csv("../data/3A4_processed/dat_train.csv") +dat_test= pd.read_csv("../data/3A4_processed/dat_test_rs.csv") + +# Hyperparameters for DNN-multitask outputs +n_out = 50 +p_missing = 0.6 + +# Preparing Data for DNN-multitask model training and testing +from utils.preprocessing import * +X_train, X_cal, y_train_norm, y_cal_norm, mu_tr, sd_tr,min_value,df_label_train, df_label_cal = preprocessing_DNN_multitask_train(dat_train, p_cal = 0.2, n_out=50,p_missing=0.6,seed=99) +X_test, df_label_test = preprocessing_DNN_default_test(dat_test, mu_tr, sd_tr) + +# Load functions for DNN-multitask model, and specify Hyperparameters +from models.DNN_multitask import * + +# Hyperparameters for DNN structure and training +p_batchSize = 0.05 +learn_rate = 0.001 + +nn_pars = {'nodes' : [4000, 2000, 1000, 1000], + 'dropout': [0.25, 0.25, 0.25, 0.1], + 'batch_size' : min(128,round(X_train.shape[0]*p_batchSize)), + 'learn_rate' : learn_rate, + 'epochs' : 500, + 'n_out': n_out, + 'min_value': min_value, + 'wt_decay' : 0.00005} + +# Train a DNN-multitask model +model = train_DNN_multitask(X_train, y_train_norm, X_cal, y_cal_norm, nn_pars) + +# Prediction on Calibration set +pred_cal_mat = model.predict(X_cal) +pred_cal_avg = np.mean(pred_cal_mat, 1) * sd_tr + mu_tr +pred_cal_sd = np.std(pred_cal_mat, 1) * sd_tr +df_pred_cal = pd.concat([df_label_cal.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_cal_avg, "Pred_UNC": pred_cal_sd})], axis=1, ignore_index = False, sort = False) + +# Calibration step +# Conformal algorithm options: CP_ACE, CP_expSD, CP_homo +from calibrators.ICP import * +nominal_level=0.8 +fun_PI = CP_ACE(df_pred_cal, nominal_level) + +# Save the model to .h5, prediction for calibration set (including raw unertainty score) to .csv, and the calibration results to .pkl file +model_path = os.path.join(saveFolder, 'model.h5') +model.save(model_path) +df_pred_cal.to_csv(os.path.join(saveFolder,"df_pred_cal.csv"), header=True, index=False) +with open(os.path.join(saveFolder, 'calibration.pkl'), 'wb') as file: + dill.dump([fun_PI,nominal_level,mu_tr,sd_tr], file) + +# Application on Test set and save results as CSV +""" +from tensorflow.keras.models import load_model +model = load_model(os.path.join(saveFolder, 'model.h5')) + +with open(os.path.join(saveFolder, 'calibration.pkl'), 'rb') as file: + fun_PI,nominal_level,mu_tr,sd_tr = dill.load(file) +""" + +pred_test_mat = model.predict(X_test) +pred_test_avg = np.mean(pred_test_mat, 1) * sd_tr + mu_tr +pred_test_sd = np.std(pred_test_mat, 1) * sd_tr +df_pred_test = pd.concat([df_label_test.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_test_avg,"Pred_UNC": pred_test_sd})], axis=1, ignore_index = False, sort = False) +df_pred_test = fun_PI(df_pred_test) +df_pred_test.to_csv(os.path.join(saveFolder,"df_pred_test.csv"), header=True, index=False) diff --git a/examples/run_LGB-tail.py b/examples/run_LGB-tail.py new file mode 100755 index 0000000..fecf114 --- /dev/null +++ b/examples/run_LGB-tail.py @@ -0,0 +1,97 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +#!/usr/bin/env python + +# python run_LGB-tail.py + +# Prerequisites +import os +import sys +import time +import glob +import numpy as np +import pandas as pd +sys.path.append("../puqsar") + +# Create folder to save results +saveFolder = '../results/LGB-tail' +if not os.path.exists(saveFolder): + os.makedirs(saveFolder) + +# Load Example Data +dat_train = pd.read_csv("../data/3A4_processed/dat_train.csv") +dat_test= pd.read_csv("../data/3A4_processed/dat_test_rs.csv") + +# Prepare Data for LGB-tail model training and testing +from utils.preprocessing import * +X_train, X_cal, y_train, y_cal, df_label_train, df_label_cal = preprocessing_default_train(dat_train, p_cal = 0.2, seed = 666) +X_test, df_label_test = preprocessing_default_test(dat_test) + +from scipy import sparse +import lightgbm as lgb +train_xy = lgb.Dataset(sparse.csr_matrix(X_train), label=y_train) + +# Train a LGB model +#import lightgbm as lgb + +param = {"num_leaves": 64, + "objective": "regression", + "metric": "mse", + "bagging_freq": 1, + "bagging_fraction": 0.7, + "feature_fraction": 0.7, + "learning_rate": 0.01, + "num_iterations": 1500, + "random_state": 1357, + "boosting_type": 'gbdt', + } +model = lgb.train(param, train_xy) + +# Prediction on Calibration set +from models.LGB_tail import * +pred_cal, pred_cal_sd = lgb_tail_preds(model, X_cal, w=0.2) +df_pred_cal = pd.concat([df_label_cal.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_cal, "Pred_UNC": pred_cal_sd})], axis=1, ignore_index = False, sort = False) + +# Calibration step +# Conformal algorithm options: CP_ACE, CP_expSD, CP_homo +from calibrators.ICP import * +nominal_level=0.8 +fun_PI = CP_ACE(df_pred_cal, nominal_level) + +# Save the model to .txt file, the calibration results to .pkl file, and the prediction for calibration set (including raw unertainty score) to .csv +import dill +model.save_model(os.path.join(saveFolder, 'model.txt')) +df_pred_cal.to_csv(os.path.join(saveFolder,"df_pred_cal.csv"), header=True, index=False) +with open(os.path.join(saveFolder, 'calibration.pkl'), 'wb') as file: + dill.dump([fun_PI,nominal_level], file) + +# Application on Test set +""" +model = lgb.Booster(model_file=os.path.join(saveFolder, 'model.txt')) + +with open(os.path.join(saveFolder, 'calibration.pkl'), 'rb') as file: + fun_PI,nominal_level = dill.load(file) +""" + +pred_test, pred_test_sd = lgb_tail_preds(model, X_test, w=0.2) +df_pred_test = pd.concat([df_label_test.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_test,"Pred_UNC": pred_test_sd})], axis=1, ignore_index = False, sort = False) +df_pred_test = fun_PI(df_pred_test) +df_pred_test.to_csv(os.path.join(saveFolder,"df_pred_test.csv"), header=True, index=False) diff --git a/examples/run_RF-OOB.py b/examples/run_RF-OOB.py new file mode 100755 index 0000000..ae08570 --- /dev/null +++ b/examples/run_RF-OOB.py @@ -0,0 +1,91 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +#!/usr/bin/env python + +# python run_RF-OOB.py + +# Prerequisites +import os +import sys +import time +import glob +import numpy as np +import pandas as pd +import dill +sys.path.append("../puqsar") + +# Create folder to save results +saveFolder = '../results/RF-OOB' +if not os.path.exists(saveFolder): + os.makedirs(saveFolder) + +# Load Example Data +dat_train = pd.read_csv("../data/3A4_processed/dat_train.csv") +dat_test= pd.read_csv("../data/3A4_processed/dat_test_rs.csv") + +# Prepare Data for RF_OOB model training and testing +from utils.preprocessing import * +X_train, y_train, df_label_train = preprocessing_RF(dat_train, returnAct = True) +X_test, df_label_test = preprocessing_RF(dat_test, returnAct = False) + +# Find number of CPUs +import multiprocessing +nCores = multiprocessing.cpu_count() + +# Train a RF model +from sklearn.ensemble import RandomForestRegressor +ntree = 500 +model = RandomForestRegressor(n_estimators=ntree, random_state=9876, max_features=0.33, min_samples_leaf=5, oob_score=True, n_jobs=nCores) +model.fit(X_train, y_train) + +# Prediction on Calibration set +from models.RF_OOB import * +pred_oob_avg, pred_oob_sd = rf_oob_preds(model, X_train) +df_pred_cal = pd.concat([df_label_train.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_oob_avg, "Pred_UNC": pred_oob_sd})], axis=1, ignore_index = False, sort = False) + +# Calibration step +# Conformal algorithm options: CP_ACE, CP_expSD, CP_homo +from calibrators.ICP import * +nominal_level=0.8 +fun_PI = CP_ACE(df_pred_cal, nominal_level) + +# Save the model and the calibration results to .pkl files, and save the prediction for calibration set (including raw unertainty score) to .csv +with open(os.path.join(saveFolder, 'model.pkl'), 'wb') as file: + dill.dump([model], file) + +df_pred_cal.to_csv(os.path.join(saveFolder,"df_pred_cal.csv"), header=True, index=False) + +with open(os.path.join(saveFolder, 'calibration.pkl'), 'wb') as file: + dill.dump([fun_PI,nominal_level], file) + +# Application on Test set and save results as CSV +""" +with open(os.path.join(saveFolder, 'model.pkl'), 'rb') as file: + model = dill.load(file) + +with open(os.path.join(saveFolder, 'calibration.pkl'), 'rb') as file: + fun_PI,nominal_level = dill.load(file) +""" + +pred_test_avg, pred_test_sd = rf_test_preds(model, X_test) +df_pred_test = pd.concat([df_label_test.reset_index(drop=True, inplace=False), + pd.DataFrame({"Pred": pred_test_avg,"Pred_UNC": pred_test_sd})], axis=1, ignore_index = False, sort = False) +df_pred_test = fun_PI(df_pred_test) +df_pred_test.to_csv(os.path.join(saveFolder,"df_pred_test.csv"), header=True, index=False) diff --git a/puqsar/__init__.py b/puqsar/__init__.py new file mode 100755 index 0000000..e69de29 diff --git a/puqsar/calibrators/ICP.py b/puqsar/calibrators/ICP.py new file mode 100755 index 0000000..611a04e --- /dev/null +++ b/puqsar/calibrators/ICP.py @@ -0,0 +1,139 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +import sys +import os +import time +import pandas as pd +import numpy as np +import math + +def CP_homo(pred_val, alpha, act_col='Act',pred_col = 'Pred', unc = 'Pred_UNC'): + absErr_val = abs(pred_val[pred_col]-pred_val[act_col]) + PI_width = np.quantile(absErr_val, alpha) # scalar + def fun_PI(pred_test, pred_col = 'Pred', unc = 'Pred_UNC', act_col='Act'): + pred_test_withPI = pred_test.copy() + pred_test_withPI['PI_width'] = PI_width + if (act_col is not None) and (act_col in list(pred_test)): + absErr_test = abs(pred_test[pred_col]-pred_test[act_col]) + test_coverage = (absErr_test <= PI_width) # vector + pred_test_withPI['testCoverage'] = test_coverage + return pred_test_withPI + return fun_PI + +def CP_expSD(pred_val, alpha, act_col='Act',pred_col = 'Pred', unc = 'Pred_UNC'): + absErr_val = abs(pred_val[pred_col]-pred_val[act_col]) + nonconformity_val = absErr_val/np.exp(pred_val[unc]) + alpha_CL = np.quantile(absErr_val, alpha) + def fun_PI(pred_test, pred_col = 'Pred', unc = 'Pred_UNC', act_col='Act'): + pred_test_withPI = pred_test.copy() + PI_width = alpha_CL*np.exp(pred_test[unc]) # vector + pred_test_withPI['PI_width'] = PI_width + if (act_col is not None) and (act_col in list(pred_test)): + absErr_test = abs(pred_test[pred_col]-pred_test[act_col]) + test_coverage = (absErr_test <= PI_width) + pred_test_withPI['testCoverage'] = test_coverage + return pred_test_withPI + return fun_PI + +def CP_ACE(pred_val, alpha, act_col='Act',pred_col = 'Pred', unc = 'Pred_UNC'): + absErr_val = abs(pred_val[pred_col]-pred_val[act_col]) + f_PredSD = fun_scaledUS(pred_val,alpha, act_col,pred_col,unc) + Pred_UNC_val = [f_PredSD(x) for x in pred_val[unc]] + nonconformity_val = absErr_val/Pred_UNC_val + alpha_CL = np.quantile(nonconformity_val, alpha) + def fun_PI(pred_test, pred_col = 'Pred', unc = 'Pred_UNC', act_col='Act'): + # Add a column of PI_width to test set + # If act_col exists, add a column of test_coverage (TRUE/FALSE) + pred_test_withPI = pred_test.copy() + Pred_UNC_test = [f_PredSD(x) for x in pred_test[unc]] + PI_width = [alpha_CL * w for w in Pred_UNC_test] + pred_test_withPI['PI_width'] = PI_width + if (act_col is not None) and (act_col in list(pred_test)): + absErr_test = abs(pred_test[pred_col]-pred_test[act_col]) + test_coverage = (absErr_test <= PI_width) + pred_test_withPI['testCoverage'] = test_coverage + return pred_test_withPI + return fun_PI + +def fun_scaledUS(pred_val, alpha, act_col='Act',pred_col = 'Pred', unc = 'Pred_UNC'): + absErr_val_all = abs(pred_val[pred_col]-pred_val[act_col]) + mu_absErr_val = np.mean(absErr_val_all) + sd_absErr_val = np.std(absErr_val_all) + mu_predSD_val = np.mean(pred_val[unc]) + sd_predSD_val = np.std(pred_val[unc]) + b = mu_absErr_val + a_max = b/sd_absErr_val + R = 20 + f = np.repeat([0,1], np.ceil(pred_val.shape[0]/2), axis = 0)[0:pred_val.shape[0]] + df_para_all = [] + for rrr in range(R): + np.random.shuffle(f) + pred_val_1 = pred_val[f==0] + pred_val_2 = pred_val[f==1] + df_k = fun_df_para(pred_val_1,pred_val_2, a_max, b, sd_absErr_val, mu_predSD_val, sd_predSD_val, alpha, act_col,pred_col,unc) + df_para_all.append(df_k) + df_para_all = pd.concat(df_para_all, axis = 0, ignore_index = True) + df_para_all.dropna(axis = 0, how = 'any', inplace = True) + df_summary_avg = df_para_all.groupby('a',as_index=True).mean().rename(columns={"test_coverage": "avg_test_coverage", + "PI_width": "avg_PI_width", + "test_coverage_errSubGroup": "avg_test_coverage_errSubGroup"}) + df_summary_sd = df_para_all.groupby('a',as_index=True).std().rename(columns={"test_coverage": "sd_test_coverage", + "PI_width": "sd_PI_width", + "test_coverage_errSubGroup": "sd_test_coverage_errSubGroup"}) + df_para_all = df_summary_avg.join(df_summary_sd, on = 'a', how = 'inner') + df_para_all.reset_index(inplace=True) + df_para_all.sort_values(by='avg_test_coverage_errSubGroup',axis = 0, ascending =True, inplace = True) + df_para_all.reset_index(inplace=True, drop = True) + threshold1 = df_para_all['avg_test_coverage_errSubGroup'][0]+df_para_all['sd_test_coverage_errSubGroup'][0]/np.sqrt(R) + df_para_subset = df_para_all[df_para_all['avg_test_coverage_errSubGroup'] <= threshold1].copy() + if df_para_subset.shape[0]>1: + df_para_subset.sort_values(by='avg_PI_width',axis = 0, ascending =True, inplace = True) + df_para_subset.reset_index(inplace=True) + threshold2 = df_para_subset['avg_PI_width'][0]+df_para_subset['sd_PI_width'][0]/np.sqrt(R) + df_para_subset = df_para_subset[df_para_subset['avg_PI_width'] <= threshold2] + a_opt = np.median(df_para_subset['a'])*sd_absErr_val + def f_PredSD(x): + return(a_opt*np.tanh((x - mu_predSD_val)/sd_predSD_val)+b) + return(f_PredSD) + + +def fun_df_para(pred_val_1, pred_val_2, a_max, b, sd_absErr_val, mu_predSD_val, sd_predSD_val, alpha, act_col='Act',pred_col = 'Pred', unc = 'Pred_UNC'): + absErr_val = abs(pred_val_1[pred_col]-pred_val_1[act_col]) + absErr_test = abs(pred_val_2[pred_col]-pred_val_2[act_col]) + n_test = len(absErr_test) + if n_test <= 2000: + subGroup=pd.cut(x=pred_val_2[pred_col], bins=np.quantile(pred_val_2[pred_col],[0,0.25,0.5,0.75,1]),labels = [1,2,3,4]) + else: + n_bins = math.ceil(n_test/500.0) + subGroup=pd.cut(x=pred_val_2[pred_col], bins=np.quantile(pred_val_2[pred_col],[0]+[i/n_bins for i in list(range(1,1+n_bins))]),labels = list(range(1,1+n_bins))) + df_para = [] + for k in range(100): + a = (k/100.0)*a_max*sd_absErr_val + Pred_UNC_val = a*np.tanh((pred_val_1[unc] - mu_predSD_val)/sd_predSD_val)+b + Pred_UNC_test = a*np.tanh((pred_val_2[unc] - mu_predSD_val)/sd_predSD_val)+b + nonconformity_val = absErr_val/Pred_UNC_val + alpha_CL = np.quantile(nonconformity_val, alpha) + PI_width = alpha_CL * Pred_UNC_test + test_coverage = (absErr_test <= PI_width) + test_coverage_k = np.mean(test_coverage) + PI_width_k = np.mean(PI_width) + df_temp = pd.DataFrame({'subGroup':subGroup, 'test_coverage': test_coverage}) + test_coverage_errSubGroup_k = np.mean(np.abs(df_temp.groupby('subGroup').mean().to_numpy()-alpha)) + df_para.append([(k/100.0)*a_max, test_coverage_k, PI_width_k, test_coverage_errSubGroup_k]) + df_para = pd.DataFrame(df_para, columns =['a', 'test_coverage', 'PI_width','test_coverage_errSubGroup']) + return(df_para) diff --git a/puqsar/calibrators/__init__.py b/puqsar/calibrators/__init__.py new file mode 100755 index 0000000..e69de29 diff --git a/puqsar/models/DNN_dropout.py b/puqsar/models/DNN_dropout.py new file mode 100755 index 0000000..1c0ef49 --- /dev/null +++ b/puqsar/models/DNN_dropout.py @@ -0,0 +1,55 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +# DNN (Fully Connected Neural Network) model +# Use random dropout during test time to compute the raw prediction uncertainty scores +# Reference: Cortes-Ciriano, I.; Bender, A. Reliable prediction errors for deep neural networks using test-time dropout. Journal of chemical information and modeling 2019, 59, 3330–3339. + +import tensorflow as tf +from tensorflow import keras +from tensorflow.keras import layers +from tensorflow.keras.metrics import RootMeanSquaredError + +def train_DNN_dropout(x, y, x_val, y_val, pars): + inputs = keras.Input(shape=(x.shape[1], )) + dense = layers.Dense(pars['nodes'][0], + activation="relu", + kernel_regularizer=keras.regularizers.l2(pars['wt_decay'])) + prv_layer = dense(inputs) + prv_layer = layers.Dropout(pars['dropout'][0])(prv_layer,training=True) + for i in range(1, len(pars['nodes'])): + prv_layer = layers.Dense(pars['nodes'][i], + activation="relu", + kernel_regularizer=keras.regularizers.l2(pars['wt_decay']))(prv_layer) + prv_layer = layers.Dropout(pars['dropout'][i])(prv_layer,training=True) + outputs = layers.Dense(1)(prv_layer) + model = keras.Model(inputs=inputs, outputs=outputs, name="DNN_dropout") + callback_ES = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=50) + model.compile( + loss=keras.losses.MeanSquaredError(), + optimizer=tf.keras.optimizers.SGD(learning_rate=pars['learn_rate'], momentum=0.9, nesterov=False), + metrics=[RootMeanSquaredError()] + ) + model.fit(x, y, + batch_size=pars['batch_size'], + epochs=pars['epochs'], + validation_data=(x_val, y_val), + shuffle=True, + callbacks=[callback_ES] + ) + return model diff --git a/puqsar/models/DNN_multitask.py b/puqsar/models/DNN_multitask.py new file mode 100755 index 0000000..9065c18 --- /dev/null +++ b/puqsar/models/DNN_multitask.py @@ -0,0 +1,62 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +# DNN (Fully Connected Neural Network) model +# Use multitask method to compute the raw prediction uncertainty scores +# Reference: Xu, Y., Liaw, A., Sheridan, R.P. and Svetnik, V., 2023. Development and Evaluation of Conformal Prediction Methods for QSAR. arXiv preprint arXiv:2304.00970. + +import tensorflow as tf +from tensorflow import keras +from tensorflow.keras import layers +from tensorflow.keras.metrics import RootMeanSquaredError +from keras import backend as K + +def masked_loss_function(y_true, y_pred): + mask = K.cast(K.greater_equal(y_true, min_value), K.floatx()) # mask for non-missing elements + return tf.keras.losses.mean_squared_error(y_true * mask, y_pred * mask) + +def train_DNN_multitask(x, y, x_val, y_val, pars): + global min_value + min_value = pars['min_value'] + inputs = keras.Input(shape=(x.shape[1], )) + dense = layers.Dense(pars['nodes'][0], + activation="relu", + kernel_regularizer=keras.regularizers.l2(pars['wt_decay'])) + prv_layer = dense(inputs) + prv_layer = layers.Dropout(pars['dropout'][0])(prv_layer) + for i in range(1, len(pars['nodes'])): + prv_layer = layers.Dense(pars['nodes'][i], + activation="relu", + kernel_regularizer=keras.regularizers.l2(pars['wt_decay']))(prv_layer) + prv_layer = layers.Dropout(pars['dropout'][i])(prv_layer) + outputs = layers.Dense(pars['n_out'])(prv_layer) + model = keras.Model(inputs=inputs, outputs=outputs, name="DNN_multitask") + callback_ES = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=50) + model.compile( + loss=masked_loss_function, + optimizer=tf.keras.optimizers.SGD(learning_rate=pars['learn_rate'], momentum=0.9, nesterov=False), + metrics=[RootMeanSquaredError()] + ) + model.fit(x, y, + batch_size=pars['batch_size'], + epochs=pars['epochs'], + validation_data=(x_val, y_val), + shuffle=True, + callbacks=[callback_ES] + ) + return model diff --git a/puqsar/models/LGB_tail.py b/puqsar/models/LGB_tail.py new file mode 100755 index 0000000..e89b6c6 --- /dev/null +++ b/puqsar/models/LGB_tail.py @@ -0,0 +1,39 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +# Companion functions for making predictions using LGB-tail method +# Reference: Xu, Y., Liaw, A., Sheridan, R.P. and Svetnik, V., 2023. Development and Evaluation of Conformal Prediction Methods for QSAR. arXiv preprint arXiv:2304.00970. + +import numpy as np + +def lgb_tail_preds(model, X, w=0.2, returnAll = False): + pred = model.predict(X) + nrounds = model.num_trees() + n_tail = int(nrounds * w / 100) + if returnAll: + pred_mat = np.zeros((X.shape[0],nrounds)) + for k in range(nrounds): + pred_mat[:,k] = model.predict(X,start_iteration=k,num_iteration=1) + pred_sd = np.mean(abs(pred_mat[:,(nrounds-n_tail):nrounds]), axis = 1) + return (pred, pred_sd, pred_mat) + else: + pred_mat = np.zeros((X.shape[0],n_tail)) + for k in range(n_tail): + pred_mat[:,k] = model.predict(X,start_iteration=(nrounds-n_tail+k),num_iteration=1) + pred_sd = np.mean(abs(pred_mat), axis = 1) + return (pred, pred_sd) diff --git a/puqsar/models/RF_OOB.py b/puqsar/models/RF_OOB.py new file mode 100755 index 0000000..70ea2b1 --- /dev/null +++ b/puqsar/models/RF_OOB.py @@ -0,0 +1,66 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +# Companion functions for using out-of-bag predictions in Random Forest model + +import numpy as np +#from sklearn.ensemble import RandomForestRegressor +from sklearn.ensemble._forest import _get_n_samples_bootstrap, _generate_unsampled_indices + +def rf_oob_preds(rf_model, X, returnTrees=False): + n_samples = X.shape[0] + oob_predictions = np.zeros((n_samples, rf_model.n_estimators)) + oob_mask = np.zeros((n_samples,rf_model.n_estimators), dtype = np.int64) + n_predictions = np.zeros((n_samples,), dtype = np.int64) + n_samples_bootstrap = _get_n_samples_bootstrap(n_samples, rf_model.max_samples) + for k, estimator in enumerate(rf_model.estimators_): + unsampled_indices = _generate_unsampled_indices(estimator.random_state, n_samples, n_samples_bootstrap) + p_estimator = estimator.predict(X[unsampled_indices, :], check_input=False) + oob_predictions[unsampled_indices, k] = p_estimator + oob_mask[unsampled_indices, k] = 1 + n_predictions[unsampled_indices] += 1 + assert((np.sum(oob_mask, axis = 1)==n_predictions).all()) + if (n_predictions == 0).any(): + warn("Some inputs do not have OOB scores. ") + temp = np.copy(n_predictions) + temp[temp == 0] = 1 + pred_oob_avg = np.sum(oob_predictions, axis = 1)/temp + else: + pred_oob_avg = np.sum(oob_predictions, axis = 1)/n_predictions + assert(np.max(np.abs(pred_oob_avg - rf_model.oob_prediction_)) < 10**(-6)) + pred_oob_sd = np.zeros(pred_oob_avg.shape) + for i in range(oob_predictions.shape[0]): + if (n_predictions[i] > 1): + pred_oob_sd[i] = np.std(oob_predictions[i,oob_mask[i,:]>0]) + else: + pred_oob_sd[i] = 0.0 + if returnTrees: + return (pred_oob_avg, pred_oob_sd, oob_predictions, oob_mask, n_predictions) + else: + return (pred_oob_avg, pred_oob_sd) + +def rf_test_preds(rf_model, X_test, returnTrees=False): + pred_test_trees = np.zeros((X_test.shape[0],rf_model.n_estimators),dtype = np.float32) + for k, tree in enumerate(rf_model.estimators_): + pred_test_trees[:,k] = tree.predict(X_test) + pred_test_avg = np.mean(pred_test_trees, axis = 1) + pred_test_sd = np.std(pred_test_trees, axis = 1) + if returnTrees: + return (pred_test_avg, pred_test_sd, pred_test_trees) + else: + return (pred_test_avg, pred_test_sd) diff --git a/puqsar/models/__init__.py b/puqsar/models/__init__.py new file mode 100755 index 0000000..e69de29 diff --git a/puqsar/utils/__init__.py b/puqsar/utils/__init__.py new file mode 100755 index 0000000..e69de29 diff --git a/puqsar/utils/preprocessing.py b/puqsar/utils/preprocessing.py new file mode 100755 index 0000000..24ca147 --- /dev/null +++ b/puqsar/utils/preprocessing.py @@ -0,0 +1,150 @@ +# Copyright © 2023 Merck & Co., Inc., Rahway, NJ, USA and its affiliates. All rights reserved. +# +# This file is part of the PUQSAR package, an open source software for computing the Prediction Uncertainty for QSAR. +# +# Prediction Uncertainty for QSAR (PUQSAR) is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + + +import sys +import os +import time +import math +import numpy as np +import pandas as pd + +""" +Prepare data for DNN_dropout or any single-task DNN model: + Create random splits of Proper Training and Calibration sets to prepare for conformal prediction + Normalize output labels + Transform input descriptors +""" +def preprocessing_DNN_default_train(dat_train, p_cal = 0.2, seed = 1234): + df_label = dat_train[['MOLECULE', 'Act']] + X = dat_train.drop(labels=['MOLECULE', 'Act'], axis = 1, inplace = False).to_numpy(dtype = "uint8") + np.random.seed(seed) + # Random split the training set into (1-p_cal) proper training and p_cal calibration set + cal_index = np.random.choice(dat_train.shape[0],size = round(p_cal*dat_train.shape[0]), replace = False) + train_index = np.setdiff1d(range(dat_train.shape[0]), cal_index) + # + X_train = X[train_index,] + X_cal = X[cal_index,] + df_label_train = df_label.iloc[train_index] + df_label_cal = df_label.iloc[cal_index] + # transform input features + X_train = np.log(X_train + 1.0).astype(np.float32) + X_cal = np.log(X_cal + 1.0).astype(np.float32) + # + y_train = df_label_train['Act'].to_numpy(dtype = np.float64) + y_cal = df_label_cal['Act'].to_numpy(dtype = np.float64) + # normalize true label + mu_tr = y_train.mean() + sd_tr = y_train.std() + y_train_norm = (y_train-mu_tr)/sd_tr + y_cal_norm = (y_cal-mu_tr)/sd_tr + return (X_train, X_cal, y_train_norm, y_cal_norm, mu_tr, sd_tr, df_label_train, df_label_cal) + +def preprocessing_DNN_default_test(dat_test, mu_tr, sd_tr): + if 'Act' in list(dat_test): + df_label_test = dat_test[['MOLECULE', 'Act']] + X_test = dat_test.drop(labels=['MOLECULE', 'Act'], axis = 1, inplace = False).to_numpy(dtype = "uint8") + else: + df_label_test = dat_test[['MOLECULE']] + X_test = dat_test.drop(labels=['MOLECULE'], axis = 1, inplace = False).to_numpy(dtype = "uint8") + # transform input features + X_test = np.log(X_test + 1.0).astype(np.float32) + return X_test, df_label_test + +""" +Prepare data for DNN_multitask model: + Create random splits of Proper Training and Calibration sets to prepare for conformal prediction + Normalize output labels + Transform input descriptors + Create multioutputs training set label with random missing for training +""" +def preprocessing_DNN_multitask_train(dat_train, p_cal = 0.2, n_out=50, p_missing=0.6,seed=1234): + df_label = dat_train[['MOLECULE', 'Act']] + X = dat_train.drop(labels=['MOLECULE', 'Act'], axis = 1, inplace = False).to_numpy(dtype = "uint8") + np.random.seed(seed) + # Random split the training set into (1-p_cal) proper training and p_cal calibration set + cal_index = np.random.choice(dat_train.shape[0],size = round(p_cal*dat_train.shape[0]), replace = False) + train_index = np.setdiff1d(range(dat_train.shape[0]), cal_index) + # + X_train = X[train_index,] + X_cal = X[cal_index,] + df_label_train = df_label.iloc[train_index] + df_label_cal = df_label.iloc[cal_index] + # transform input features + X_train = np.log(X_train + 1.0).astype(np.float32) + X_cal = np.log(X_cal + 1.0).astype(np.float32) + # + y_train = df_label_train['Act'].to_numpy(dtype = np.float64) + y_cal = df_label_cal['Act'].to_numpy(dtype = np.float64) + # normalize true label + mu_tr = y_train.mean() + sd_tr = y_train.std() + y_train_norm = (y_train-mu_tr)/sd_tr + y_cal_norm = (y_cal-mu_tr)/sd_tr + # multitask output with missing + min_value = np.min(y_train_norm) + mask_value = min_value-1000.0 + y_train_mat = np.tile(y_train_norm,(n_out,1)).transpose() + y_cal_mat = np.tile(y_cal_norm,(n_out,1)).transpose() + y_train_mask = np.random.choice([0,1], size = y_train_mat.shape, p=[p_missing,(1-p_missing)]) + y_train_missing = y_train_mat*y_train_mask + mask_value*(1-y_train_mask) + return (X_train, X_cal, y_train_missing, y_cal_mat, mu_tr, sd_tr, min_value, df_label_train, df_label_cal) + +""" +Prepare data for RF_OOB model: + Since OOB data will be used for calibration, there is no need of splitting or preprocessing for both training and test sets. +""" +def preprocessing_RF(dat, returnAct = False): + if 'Act' in list(dat): + df_label = dat[['MOLECULE', 'Act']] + X = dat.drop(labels=['MOLECULE', 'Act'], axis = 1, inplace = False).to_numpy(dtype = np.float32) + if returnAct: + y = df_label['Act'].to_numpy(dtype = np.float64) + return X, y, df_label + else: + df_label = dat[['MOLECULE']] + X = dat.drop(labels=['MOLECULE'], axis = 1, inplace = False).to_numpy(dtype = np.float32) + return X, df_label + +""" +Prepare data for LGB_tail or other general ML models: + Create random splits of Proper Training and Calibration sets to prepare for conformal prediction +""" +def preprocessing_default_train(dat_train, p_cal = 0.2, seed = 1234): + df_label = dat_train[['MOLECULE', 'Act']] + X = dat_train.drop(labels=['MOLECULE', 'Act'], axis = 1, inplace = False).to_numpy(dtype = "uint8") + np.random.seed(seed) + # Random split the training set into (1-p_cal) proper training and p_cal calibration set + cal_index = np.random.choice(dat_train.shape[0],size = round(p_cal*dat_train.shape[0]), replace = False) + train_index = np.setdiff1d(range(dat_train.shape[0]), cal_index) + # + X_train = X[train_index,].astype(np.float32) + X_cal = X[cal_index,].astype(np.float32) + df_label_train = df_label.iloc[train_index] + df_label_cal = df_label.iloc[cal_index] + y_train = df_label_train['Act'].to_numpy(dtype = np.float64) + y_cal = df_label_cal['Act'].to_numpy(dtype = np.float64) + return (X_train, X_cal, y_train, y_cal, df_label_train, df_label_cal) + +def preprocessing_default_test(dat_test): + if 'Act' in list(dat_test): + df_label_test = dat_test[['MOLECULE', 'Act']] + X_test = dat_test.drop(labels=['MOLECULE', 'Act'], axis = 1, inplace = False).to_numpy(dtype = np.float32) + else: + df_label_test = dat_test[['MOLECULE']] + X_test = dat_test.drop(labels=['MOLECULE'], axis = 1, inplace = False).to_numpy(dtype = np.float32) + return X_test, df_label_test