Spaces:
Sleeping
Sleeping
Mohamudzuly
commited on
Commit
•
53e2810
1
Parent(s):
1a95597
Create code_structure_documentation.md
Browse files
innovation_pathfinder_ai/structured_tools/code_structure_documentation.md
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
## Functionality Breakdown
|
2 |
+
|
3 |
+
### Internal Knowledge Exploration
|
4 |
+
|
5 |
+
- **memory_search**: Searches a Chroma vector store containing past findings, utilizing sentence embeddings to identify relevant information.
|
6 |
+
|
7 |
+
- **knowledgeBase_search**: Similar to `memory_search`, but specifically explores a collection within Chroma potentially holding research papers or relevant information chunks.
|
8 |
+
|
9 |
+
### External Information Retrieval
|
10 |
+
|
11 |
+
- **arxiv_search**: Retrieves scientific research papers from the Arxiv database based on user queries, stores metadata, and formats papers for display.
|
12 |
+
|
13 |
+
- **get_arxiv_paper**: Downloads full PDF content of a specific Arxiv paper based on its ID and saves it to a designated directory.
|
14 |
+
|
15 |
+
- **wikipedia_search**: Queries Wikipedia for relevant summaries based on user input, formats and stores retrieved summaries.
|
16 |
+
|
17 |
+
- **google_search**: Performs a Google Search using user queries, retrieves relevant results, and stores them for later use.
|
18 |
+
|
19 |
+
### Data Storage and Management
|
20 |
+
|
21 |
+
- **Chroma Vector Store**: Stores text content embeddings for similarity search (`memory_search`) and Arxiv paper embeddings (`embed_arvix_paper`).
|
22 |
+
|
23 |
+
- **Global `all_sources` Variable**: Accumulates retrieved information from various sources (Arxiv, Wikipedia, Google Search).
|
24 |
+
|
25 |
+
## Additional Notes
|
26 |
+
|
27 |
+
- These tools are integrated with a larger LangChain framework, as indicated by the `@tool` decorator.
|
28 |
+
|
29 |
+
- Configuration files (`config.ini`) define details like vector store location and collection names.
|
30 |
+
|
31 |
+
## Getting Started
|
32 |
+
|
33 |
+
1. Install the LangChain framework and required libraries.
|
34 |
+
|
35 |
+
2. Configure access to the Chroma vector store (likely involves environment variables or configuration files).
|
36 |
+
|
37 |
+
3. Refer to the InnovationPathfinderAI project documentation for detailed integration instructions.
|
38 |
+
|
39 |
+
## Disclaimer
|
40 |
+
|
41 |
+
This documentation provides an overview of the tools' functionalities.
|