HydroShare on Jupyter 0.1.5

• Support for Safari is added (see issue #98 for background).
• A minor frontend server root path issue is resolved (see pr #99's summary for details)
• Both frontend and backend are bumped to 0.1.5 (resolves #97)

Additions

• Backend:
  • ServerRootHandler added. Endpoint responds with expanded absolute directory path location of where JupyterHub session was started. Endpoint path is: /syncApi/root_directory/.

Changes

• Frontend:
  • Location of serverRoot is now retrieved from backend ServerRootHandler at /syncApi/root_directory/
  • No longer use Array.at to index array. Resolves #98

HydroShare on Jupyter 0.1.4

Resolves issue with configuration handler failing to create intermediate directories. (see #92)

Changes

• ConfigFile now creates intermediate directories for logging and data directories if they do not already exist.

HydroShare on Jupyter 0.1.3

Issue when downloading an entire resource from backend. Previously, resources were unzipped to data_path / resource_id. They are now, correctly, unzipped to data_path / resource_id / resource_id. Issue with resource files being shown as only local, when they are in-sync. This was likely because when a file was downloaded from hydroshare and moved to the correct file-system location, the remote fs resource map had not completed its update causing all files to appear as local.

Changes

• Fixes unpacking location of entire resource upon download.
• Fixes remote fs resource map race condition, where files might have appeared to only be on the local system.
• websocket handler no longer subscribes to resource downloaded and resource entity downloaded events. Given that the current event system dispatches callbacks sequentially in the order of subscription (works like queue), if the websocket were to have subscribed to these events before the aggregate filesystem (the aggregate filesystem updates itself based off of these events), it was possible to have a false state.

Related PR's

#91

HydroShare on Jupyter 0.1.2

Backend now lazily fills fs aggregate map as a user queries for files in a given resource. This drastically decreases start-up time post login and resolves websocket time out issues. For context, websocket timeout were increasingly likely proportional to the number of local resources.

Additions

• Add missing init.py to utilities submodule

Changes

• File system map now lazily adds local resources to the map when a resource's files are queried

Related PR's

#89, #90

HydroShare on Jupyter 0.1.0

What's Changed

• Package now named hydroshare_on_jupyter.
• Add build system deps and update setup.py metadata by @aaraney in #78
• Add pull request template by @aaraney in #79
• Re-architect backend tornado http server. Adds supporting decoupled library packages. (Large PR) by @aaraney in #80
• UI redesign and reimplimentation as jupyter lab plugin by @aaraney in #84
• Add mechanism to login to HydroShare via OAuth2. Adds backend and frontend support by @aaraney in #85
• Rename package to hydroshare_on_jupyter and cleanup package by @aaraney in #86
• Reset package version to 0.1.0 by @aaraney in #87

New Contributors

• @aaraney made their first contribution in #78

Full Changelog: v0.1.10...v0.1.0