From bfbd396c399de7ee8067dc5218e02705f9caa4bf Mon Sep 17 00:00:00 2001
From: Paco Nathan <ceteri@gmail.com>
Date: Wed, 10 Jan 2024 16:43:09 -0800
Subject: [PATCH] unit test coverage for deser

---
 docs/index.md       |  2 +-
 pkg_doc.cfg         |  3 ++-
 tests/test_load.py  | 62 +++++++++++++++++++++++++++++++++++++++++++++
 textgraphs/graph.py | 55 ++++++++++++++++++++++++++++++++++++++++
 4 files changed, 120 insertions(+), 2 deletions(-)
 create mode 100644 tests/test_load.py

diff --git a/docs/index.md b/docs/index.md
index 09c4dd7..5bad87d 100644
--- a/docs/index.md
+++ b/docs/index.md
@@ -18,7 +18,7 @@ expensive process and poor results, relative to other methods.
 
 That said, are there other ways transformers might help augment
 natural language workflows?
-This project results from an ongoing pursuit of that line of inquiry.
+This project results from an ongoing pursuit along these lines of inquiry.
 With sufficiently narrowed task focus and ample software engineering,
 transformers can be used to augment specific _components_ of natural
 language workflows.
diff --git a/pkg_doc.cfg b/pkg_doc.cfg
index 214e46d..494b821 100644
--- a/pkg_doc.cfg
+++ b/pkg_doc.cfg
@@ -7,8 +7,9 @@
 	"TextGraphs",
         "SimpleGraph",
         "Node",
-        "NodeEnum",
         "Edge",
+        "EnumBase",
+        "NodeEnum",
         "RelEnum",
         "PipelineFactory",
         "Pipeline",
diff --git a/tests/test_load.py b/tests/test_load.py
new file mode 100644
index 0000000..68ab4da
--- /dev/null
+++ b/tests/test_load.py
@@ -0,0 +1,62 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+"""
+unit tests:
+
+  * serialization and deserialization
+
+see copyright/license https://huggingface.co/spaces/DerwenAI/textgraphs/blob/main/README.md
+"""
+
+from os.path import abspath, dirname
+import json
+import pathlib
+import sys
+
+import deepdiff  # pylint: disable=E0401
+
+sys.path.insert(0, str(pathlib.Path(dirname(dirname(abspath(__file__))))))
+import textgraphs  # pylint: disable=C0413
+
+
+def test_load_minimal (
+    ) -> None:
+    """
+Construct a _lemma graph_ from a minimal example, then compare
+serialized and deserialized data to ensure no fields get corrupted
+in the conversions.
+    """
+    text: str = """
+See Spot run.
+    """
+
+    tg: textgraphs.TextGraphs = textgraphs.TextGraphs()  # pylint: disable=C0103
+    pipe: textgraphs.Pipeline = tg.create_pipeline(text.strip())
+
+    # serialize into node-link format
+    tg.collect_graph_elements(pipe)
+    tg.construct_lemma_graph()
+    tg.calc_phrase_ranks()
+
+    json_str: str = tg.dump_lemma_graph()
+    exp_graph = json.loads(json_str)
+
+    # deserialize from node-link format
+    tg = textgraphs.TextGraphs()  # pylint: disable=C0103
+    tg.load_lemma_graph(json_str)
+    tg.construct_lemma_graph()
+
+    obs_graph: dict = json.loads(tg.dump_lemma_graph())
+
+    # compare
+    diff: deepdiff.diff.DeepDiff = deepdiff.DeepDiff(exp_graph, obs_graph)
+
+    if len(diff) > 0:
+        print(json.dumps(json.loads(diff.to_json()), indent = 2))
+
+    assert len(diff) == 0
+
+
+if __name__ == "__main__":
+    test_load_minimal()
diff --git a/textgraphs/graph.py b/textgraphs/graph.py
index 705935e..4ec2588 100644
--- a/textgraphs/graph.py
+++ b/textgraphs/graph.py
@@ -281,3 +281,58 @@ def dump_lemma_graph (
             indent = 2,
             separators = ( ",", ":" ),
         )
+
+
+    def load_lemma_graph (
+        self,
+        json_str: str,
+        ) -> None:
+        """
+Load from a JSON string in
+a JSON representation of the exported _lemma graph_ in
+[_node-link_](https://networkx.org/documentation/stable/reference/readwrite/json_graph.html)
+format
+        """
+        dat: dict = json.loads(json_str)
+        tokens: typing.List[ Node ] = []
+
+        # deserialize the nodes
+        for nx_node in dat.get("nodes"):  # type: ignore
+            label: typing.Optional[ str ] = None
+            kind: NodeEnum = NodeEnum.decode(nx_node["kind"])  # type: ignore
+
+            if kind in [ NodeEnum.ENT ]:
+                label = nx_node["label"]
+
+            node: Node = self.make_node(
+                tokens,
+                nx_node["lemma"],
+                None,
+                kind,
+                0,
+                0,
+                0,
+                label = label,
+                length = nx_node["length"],
+            )
+
+            node.text = nx_node["name"]
+            node.pos = nx_node["pos"]
+            node.loc = eval(nx_node["loc"])  # pylint: disable=W0123
+            node.count = int(nx_node["count"])
+            node.neighbors = int(nx_node["hood"])
+
+        # deserialize the edges
+        node_list: typing.List[ Node ] = list(self.nodes.values())
+
+        for nx_edge in dat.get("links"):  # type: ignore
+            edge: Edge = self.make_edge(  # type: ignore
+                node_list[nx_edge["source"]],
+                node_list[nx_edge["target"]],
+                RelEnum.decode(nx_edge["kind"]),  # type: ignore
+                nx_edge["title"],
+                float(nx_edge["prob"]),
+                key = nx_edge["lemma"],
+            )
+
+            edge.count = int(nx_edge["count"])