0.1.0 release

README.md

@@ -1,25 +1,14 @@
-# RAGHandler for PostgreSQL
+# RAGHandler for PostgreSQL with pgvector
 
 A **Retrieval-Augmented Generation (RAG)** handler class built around a PostgreSQL database and OpenAI’s API. This code allows you to query a database, generate vector embeddings for columns, and integrate these operations with a conversational Large Language Model (LLM), such as GPT.
 
 ## Features
 
 - **System Prompt Generation:** Automates the creation of a system prompt that describes your PostgreSQL table schema and sample rows.
 
+- **Sophisticated SQL Execution**:  Runs multi-step queries with pgvector embeddings for semantic search and filtering.
 
-- **SQL Execution & Vector Replacement**: Executes SQL queries directly against the database. If the SQL query contains <vector>TEXT<vector/> placeholders, those placeholders are replaced with vector embeddings generated via OpenAI.
-
-- **Chat-Based Interaction:** Supports a chat-based workflow by combining user messages, system prompts, and LLM responses. Integrates function calls for:
-
- 	- Executing SQL queries (execute_sql_query),
- 	- Structuring objects for potential insertion (structure_object_from_draft),
-	-	Generating Plotly graphs (plot_graph).
-
-- **Schema Summarization**: Summarizes a database table’s schema by prompting the LLM, providing a quick overview of column purposes and example values.
-
-- **Object Structuring**: Helps transform free-form text “drafts” into structured objects (e.g., JSON) for database insertion.
-
-- **Embedding Integration**: Automatically generates and stores vector embeddings for specified text columns when creating or replacing tables.
+- **Plotly Graph Generation**: Helps transform free-form text “drafts” into structured objects (e.g., JSON) for database insertion.
 
 
 ## Installation

rag_handler/__init__.py

@@ -0,0 +1,2 @@
+# __init__.py
+from .rag_handler import RAGHandler, VECTOR

setup.py

@@ -0,0 +1,17 @@
+from setuptools import setup, find_packages
+
+setup(
+    name='rag-handler',
+    version='0.1.0',
+    description='Retrieval-Augmented Generation Handler using PostgreSQL and OpenAI',
+    author='Quentin Fuxa',
+    packages=find_packages(),
+    install_requires=[
+        'openai',
+        'pandas',
+        'psycopg2',
+        'sqlalchemy',
+        'plotly'
+    ],
+    python_requires='>=3.9'
+)