dev update

.gitignore

@@ -1,3 +1,5 @@
+*.pdf
+*.json
 .jukit/
 
 # Byte-compiled / optimized / DLL files

create_common_report.py

@@ -0,0 +1,21 @@
+from fgutils.proxy_collection.common import common_groups
+from fgutils.proxy import MolProxy
+
+from fgutils.vis import PdfWriter, GraphVisualizer
+
+top_level_groups = ["alkyl", "aryl", "allyl", "alkene", "amine", "ester", "ether", "carbonyl"]
+
+vis = GraphVisualizer()
+def plot(value, ax, length=0, group="", index=0):
+    vis.plot(value, ax, title="{} {}/{}".format(group, index + 1, length))
+
+pdf_writer = PdfWriter("common_groups.pdf", plot_fn=plot, cols=3)
+
+data = common_groups
+for tlg in top_level_groups:
+    proxy = MolProxy("{{{}}}".format(tlg), common_groups)
+    data = list(proxy)
+    print("Group '{}' has {} molecules.".format(tlg, len(data)))
+    pdf_writer.plot(data, group=tlg, length=len(data))
+
+pdf_writer.close()

demo.py

@@ -1,46 +1,36 @@
 import numpy as np
 from fgutils.proxy_collection.diels_alder_proxy import DielsAlderProxy
-from fgutils.vis import GraphVisualizer, plot_reaction, plot_its
+from fgutils.vis import GraphVisualizer, plot_reaction, plot_its, PdfWriter
 from fgutils.chem.its import get_its
 
 import matplotlib.pyplot as plt
 from matplotlib.backends.backend_pdf import PdfPages
 
+
+def plot(data, ax, index=0, **kwargs):
+    g, h = data
+    its = get_its(g, h)
+    plot_reaction(g, h, ax[0])
+    ax[0].set_title("Reaction | Index: {}".format(index))
+    plot_its(its, ax[1])
+    ax[1].set_title("ITS | Index: {}".format(index))
+
+
 neg_sample = True
 
 if neg_sample:
     file_name = "DA_reactions_neg.pdf"
 else:
     file_name = "DA_reactions.pdf"
+
 proxy = DielsAlderProxy(neg_sample=neg_sample)
 vis = GraphVisualizer()
+pdf_writer = PdfWriter(
+    file_name, plot, max_pages=50, plot_per_row=True, width_ratios=[2, 1]
+)
+
+data = list(proxy)
+print("Generated {} reactions.".format(len(data)))
 
-data = [r for r in proxy]
-print("Created {} reactions.".format(len(data)))
-
-pages = 10
-rows, cols = 7, 2
-n_print = pages * rows
-step = np.max([len(data) / n_print, 1])
-pp = PdfPages(file_name)
-done = False
-for p in range(pages):
-    print("Pring page {} of {}".format(p + 1, pages))
-    fig, ax = plt.subplots(rows, cols, figsize=(21, 29.7), width_ratios=[2, 1])
-    for r in range(rows):
-        _idx = int((p * rows + r) * step)
-        if _idx >= len(data):
-            done = True
-            break
-        g, h = data[_idx]
-        its = get_its(g, h)
-        plot_reaction(g, h, ax[r, 0])
-        ax[r, 0].set_title("Reaction | Index: {}".format(_idx))
-        plot_its(its, ax[r, 1])
-        ax[r, 1].set_title("ITS | Index: {}".format(_idx))
-    plt.tight_layout()
-    pp.savefig(fig, bbox_inches="tight", pad_inches=1)
-    plt.close()
-    if done:
-        break
-pp.close()
+pdf_writer.plot(data)
+pdf_writer.close()

doc/figure_scripts/proxy_collection_DAReactionProxy.py

@@ -12,6 +12,7 @@
 from fgutils.parse import Parser
 from fgutils.proxy_collection.diels_alder_proxy import DielsAlderProxy
 
+plot = False
 
 def get_group(groups, name):
     for g in groups:
@@ -31,7 +32,8 @@ def plot_core_graphs():
         vis.plot(parser(g.pattern), ax[i], title=g.name)
     plt.tight_layout()
     plt.savefig("doc/figures/proxy_collection_DAReactionProxy.png", bbox_inches="tight")
-    plt.show()
+    if plot:
+        plt.show()
 
 
 def plot_graphs(graphs, name, cols=4, show_aam=False):
@@ -58,7 +60,8 @@ def plot_graphs(graphs, name, cols=4, show_aam=False):
         "doc/figures/proxy_collection_DAReactionProxy_{}.png".format(name),
         bbox_inches="tight",
     )
-    plt.show()
+    if plot:
+        plt.show()
 
 
 def plot_group(group, name, cols=4):
@@ -114,9 +117,9 @@ def plot_samples(neg_sample=False):
     data = [r for r in proxy]
     print("Created {} reactions.".format(len(data)))
 
-    rows, cols = 7, 2
+    rows, cols = 10, 2
     step = np.max([len(data) / rows, 1])
-    fig, ax = plt.subplots(rows, cols, figsize=(21, 29.7), width_ratios=[2, 1])
+    fig, ax = plt.subplots(rows, cols, figsize=(21, 4 * rows), width_ratios=[2, 1])
     for r in range(rows):
         _idx = int(r * step)
         if _idx >= len(data):
@@ -132,7 +135,8 @@ def plot_samples(neg_sample=False):
         "doc/figures/proxy_collection_DAReactionProxy_{}.png".format(file_name),
         bbox_inches="tight",
     )
-    plt.show()
+    if plot:
+        plt.show()
 
 
 def plot_s_trans_diene():

doc/proxy_collection.rst

@@ -70,7 +70,7 @@ parent, are surrounded by ``[]``.
 .. image:: figures/proxy_collection_DAReactionProxy_EWG.png
    :width: 1000
 
-**A few of the 2873 Diels-Alder reactions generated by the proxy:**
+**A few of the 10470 Diels-Alder reactions generated by the proxy:**
 
 .. image:: figures/proxy_collection_DAReactionProxy_DA_reactions.png
    :width: 1000
@@ -89,7 +89,8 @@ example in a fixed s-trans conformation.
 .. image:: figures/proxy_collection_DAReactionProxy_s-trans_diene.png
    :width: 1000
 
-**A few of the 6253 INVALID Diels-Alder reactions generated by the proxy (negative data):**
+**A few of the 12875 INVALID Diels-Alder reactions generated by the proxy
+(negative data):**
 
 .. image:: figures/proxy_collection_DAReactionProxy_DA_reactions_neg.png
    :width: 1000

export_DA.py

@@ -0,0 +1,22 @@
+import json
+
+from fgutils.proxy_collection import DielsAlderProxy
+from fgutils.rdkit import graph_to_smiles
+
+def export(file_name, neg_sample):
+    idx_fmt = "DA_{}"
+    if neg_sample:
+        idx_fmt = "DA_neg_{}"
+
+    dataset = []
+    proxy = DielsAlderProxy(neg_sample=neg_sample)
+    for i, r in enumerate(proxy):
+        g, h = r
+        smiles = "{}>>{}".format(graph_to_smiles(g), graph_to_smiles(h))
+        dataset.append({"index": idx_fmt.format(i), "reaction": smiles})
+
+    with open(file_name, "w") as f:
+        json.dump(dataset, f, indent=4)
+
+export("Diels-Alder_synthetic_data.json", neg_sample=False)
+export("Diels-Alder_synthetic_negative_data.json", neg_sample=True)

fgutils/parse.py

@@ -6,7 +6,7 @@
 
 
 token_specification = [
-    ("ATOM", r"H|Br|Cl|Se|Sn|Si|C|N|O|P|S|F|B|I|b|c|n|o|p|s"),
+    ("ATOM", r"H|Br|Cl|Se|Sn|Si|Mg|Li|C|N|O|P|S|F|B|I|b|c|n|o|p|s"),
     ("BOND", r"\.|-|=|#|$|:|/|\\"),
     ("BRANCH_START", r"\("),
     ("BRANCH_END", r"\)"),
@@ -116,14 +116,6 @@ def __process_token_ring(self, value):
         if value in self.rings.keys():
             anchor_sym = self.graph.nodes[self.anchor][SYMBOL_KEY]
             ring_anchor = self.rings[value]
-            ring_anchor_sym = self.graph.nodes[ring_anchor][SYMBOL_KEY]
-            if anchor_sym.islower() != ring_anchor_sym.islower():
-                raise SyntaxError(
-                    (
-                        "Ring {} must be of same aromaticity type. "
-                        + "Started with {} and ended with {}."
-                    ).format(value, ring_anchor_sym, anchor_sym)
-                )
             if anchor_sym.islower():
                 self.bond_order = 1.5
             if self.bond_order is not None:

fgutils/proxy.py

@@ -414,8 +414,12 @@ def groups(self, value):
         elif isinstance(value, list):
             for group in value:
                 self.__groups[group.name] = group
-        else:
+        elif isinstance(value, dict) and isinstance(
+            list(value.values())[0], ProxyGroup
+        ):
             self.__groups = value
+        else:
+            raise TypeError("Invalid group type.")
 
     def __str__(self):
         s = "ReactionProxy | Core: {} Enable AAM: {}\n".format(