fixes

Author: epinzur

docs/examples/index.md

@@ -18,12 +18,12 @@
     vector store so that it can be used to provide context to an LLM for code generation.
 
     It uses LangChain and `langchain-graph-retriever` with a custom traversal Strategy
-    to in order to improve LLM generated code output. It shows that using GraphRAG can
-    provide a significant increase in quality over using either a LLM alone or standard
+    in order to improve LLM generated code output. It shows that using GraphRAG can
+    provide a significant increase in quality over using either an LLM alone or standard
     RAG.
 
-    GraphRAG is traverses through the documentation in a way similar to how a
-    software engineer would, in order to determine how to solve a coding problem.
+    GraphRAG traverses cross references in the documentation like a software engineer
+    would, in order to determine how to solve a coding problem.
 
     [:material-fast-forward: Code Generation Example](code-generation.ipynb)
 </div>

docs/examples/lazy-graph-rag.ipynb

@@ -71,13 +71,6 @@
     "The last package -- `graph-rag-example-helpers` -- includes some helpers for setting up environment helpers and allowing the loading of wikipedia data to be restarted if it fails."
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  },
   {
    "cell_type": "code",
    "execution_count": null,

packages/graph-rag-example-helpers/src/graph_rag_example_helpers/datasets/astrapy/__init__.py

@@ -1,9 +1,5 @@
-from ...examples.code_generation.format import add_tabs, format_docs, format_document
 from .fetch import fetch_documents
 
 __all__ = [
     "fetch_documents",
-    "add_tabs",
-    "format_document",
-    "format_docs",
 ]

packages/graph-rag-example-helpers/src/graph_rag_example_helpers/examples/code_generation/__init__.py

@@ -1,7 +1,6 @@
-from .format import add_tabs, format_docs, format_document
+from .format import format_docs, format_document
 
 __all__ = [
-    "add_tabs",
     "format_document",
     "format_docs",
 ]

packages/graph-rag-example-helpers/src/graph_rag_example_helpers/examples/code_generation/converter.py

@@ -2,12 +2,11 @@
 import json
 import os
 import typing
+from textwrap import indent
 from typing import Any
 
 import griffe
 
-from graph_rag_example_helpers.utils import add_tabs
-
 
 def convert(
     package_name: str,
@@ -165,7 +164,7 @@ def _format_parameter(self, el: dict[str, str]) -> str:
         if "default" in el:
             text += f" = {el['default']}"
         if "description" in el:
-            desc = add_tabs(el["description"])
+            desc = indent(el["description"], "\t")
             text += f"\n\t{desc}"
         return text
 
@@ -174,7 +173,7 @@ def _format_return(self, el: dict[str, str]) -> str:
         if "type" in el:
             items.append(el["type"])
         if "description" in el:
-            items.append(add_tabs(el["description"]))
+            items.append(indent(el["description"], "\t"))
         return "\n\t".join(items)
 
     def _extract_common(self, obj: griffe.Object) -> dict[str, Any]:

packages/graph-rag-example-helpers/src/graph_rag_example_helpers/examples/code_generation/format.py

@@ -1,6 +1,10 @@
+from textwrap import indent
+
 from langchain_core.documents import Document
 
-from graph_rag_example_helpers.utils import add_tabs
+
+def _add_tabs(text: str) -> str:
+    return indent(text, "\t")
 
 
 def _format_parameter(el: dict[str, str]) -> str:
@@ -13,7 +17,7 @@ def _format_parameter(el: dict[str, str]) -> str:
     if "default" in el:
         text += f" = {el['default']}"
     if "description" in el:
-        desc = add_tabs(el["description"])
+        desc = _add_tabs(el["description"])
         text += f"\n\t{desc}"
     return text
 
@@ -23,7 +27,7 @@ def _format_return(el: dict[str, str]) -> str:
     if "type" in el:
         items.append(el["type"])
     if "description" in el:
-        items.append(add_tabs(el["description"]))
+        items.append(_add_tabs(el["description"]))
     return "\n\t".join(items)
 
 
@@ -36,30 +40,30 @@ def format_document(doc: Document, debug: bool = False) -> str:
 
     for key in ["bases", "exports", "implemented_by"]:
         if key in metadata:
-            values = '\n'.join(metadata[key])
-            text += f"{key}: \n\t{add_tabs(values)}\n\n"
+            values = "\n".join(metadata[key])
+            text += f"{key}: \n\t{_add_tabs(values)}\n\n"
 
     if "properties" in metadata:
         props = [f"{k}: {v}" for k, v in metadata["properties"].items()]
-        text += f"properties: \n\t{add_tabs('\n'.join(props))}\n\n"
+        text += f"properties: \n\t{_add_tabs('\n'.join(props))}\n\n"
 
     if doc.page_content != "":
-        text += f"description: \n\t{add_tabs(doc.page_content)}\n\n"
+        text += f"description: \n\t{_add_tabs(doc.page_content)}\n\n"
     elif "value" in metadata:
         text += f"{metadata['value']}\n\n"
 
     for key in ["attributes", "parameters"]:
         if key in metadata:
-            values = '\n\n'.join([_format_parameter(v) for v in metadata[key]])
-            text += f"{key}: \n\t{add_tabs(values)}\n\n"
+            values = "\n\n".join([_format_parameter(v) for v in metadata[key]])
+            text += f"{key}: \n\t{_add_tabs(values)}\n\n"
 
     for key in ["returns", "yields"]:
         if key in metadata:
-            values = '\n\n'.join([_format_return(v) for v in metadata[key]])
-            text += f"{key}: \n\t{add_tabs(values)}\n\n"
+            values = "\n\n".join([_format_return(v) for v in metadata[key]])
+            text += f"{key}: \n\t{_add_tabs(values)}\n\n"
 
     for key in ["note", "example"]:
-        text += f"{key}: \n\t{add_tabs(metadata[key])}\n\n"
+        text += f"{key}: \n\t{_add_tabs(metadata[key])}\n\n"
 
     if debug:
         if "imports" in metadata:
@@ -70,12 +74,12 @@ def format_document(doc: Document, debug: bool = False) -> str:
                 else:
                     imports.append(f"{real_name} as {as_name}")
             values = "\n".join(imports)
-            text += f"imports: \n\t{add_tabs(values)}\n\n"
+            text += f"imports: \n\t{_add_tabs(values)}\n\n"
 
         for key in ["references", "gathered_types"]:
             if key in metadata:
                 values = "\n".join(metadata[key])
-                text += f"{key}: \n\t{add_tabs(values)}\n\n"
+                text += f"{key}: \n\t{_add_tabs(values)}\n\n"
 
         if "parent" in metadata:
             text += f"parent: {metadata['parent']}\n\n"