feat: add ability to chat with already embedded documents

README.md

@@ -2,22 +2,29 @@
 
 <div align="center">
 
- **_lesa_**
- `[lee - saa]` • **Old Norse** <br/>
- (v.) to read, to study, to learn
+[![Python](https://img.shields.io/badge/python-3.10%2B-blue)](https://www.python.org/downloads/)
+![PyPI - Version](https://img.shields.io/pypi/v/lesa)
+![PyPI Downloads](https://static.pepy.tech/badge/lesa)
 
 </div>
 
-`lesa` is a CLI tool built in Python that allows you to converse with your documents from the terminal, completely offline and on-device using **Ollama**. Open the terminal in the directory of your choice and start a conversation with any document!
-
 <div align="center">
 
-[![Python](https://img.shields.io/badge/python-3.10%2B-blue)](https://www.python.org/downloads/)
-![PyPI - Version](https://img.shields.io/pypi/v/lesa)
-![PyPI Downloads](https://static.pepy.tech/badge/lesa)
+**_lesa_**
+`[lee - saa]` • **Old Norse** <br/>
+(v.) to read, to study, to learn
+
+<!-- <div align="center">
+  <sub>Prepared by <a href="https://github.com/shxntanu">Shantanu Wable</a> and <a href="https://github.com/omkargwagholikar">Omkar Wagholikar</a> </sub>
+</div> -->
 
 </div>
 
+`lesa` is a CLI tool built in Python that allows you to converse with your documents from the terminal, completely offline and on-device using **Ollama**. Open the terminal in the directory of your choice and start a conversation with any document!
+
+
+
+
 ## Usage
 
 To start a conversation with a document (`.pdf` and `.docx` for now), simply run:
@@ -26,7 +33,13 @@ To start a conversation with a document (`.pdf` and `.docx` for now), simply run
 lesa read path/to/your/document
 ```
 
-### Embed 
+Or start a conversation with an already-embedded directory, run:
+
+```bash
+lesa chat
+```
+
+### Embed
 
 To embed all files from your current working directory, run:
 
@@ -46,21 +59,15 @@ This creates a `.lesa` config folder in your current working directory that stor
 
 ### Prerequisites
 
-This project uses [Ollama](https://ollama.com/) under the hood to utilize the power of large language models. To install Ollama, run:
+`lesa` uses [Ollama](https://ollama.com/) under the hood to utilize the power of large language models. To install Ollama, run:
 
 ```bash
 curl -fsSL https://ollama.com/install.sh | sh
 ```
 
-This project uses the Llama 3.1 8b or Qwen 4b model as the default. You can use any other model as well, just make sure it has enough context window to understand the content of your documents.
-
-Pull Llama or Qwen using:
-
-```bash
-ollama pull qwen:4b
-```
+`lesa` uses the Llama 3.1 8b as the default. You can use any other model as well, just make sure it has enough context window to understand the content of your documents.
 
-or
+Pull Llama using:
 
 ```bash
 ollama pull llama3.1
@@ -86,4 +93,4 @@ We welcome contributions! If you'd like to improve `lesa` or have any feedback,
 
 ## License
 
-Apache-2.0
+Apache-2.0

lesa/cli/main.py

@@ -146,6 +146,19 @@ def read(file_path: str = typer.Argument(..., help="Path of the file to read")):
 
     return cm.embed_single_document_and_chat(file_path)
 
+@app.command()
+def chat():
+    """
+    Starts a chat with the embedded documents.
+    """
+
+    if not OllamaManager.is_server_running():
+        console.print(f"[red]Error: Ollama server is not running![/red]")
+        console.print(f"Start the ollama server by running [cyan]lesa start[/cyan] command.")
+        raise typer.Exit(1)
+
+    return cm.start_conversation()
+
 @app.command()
 def use(model_name: str = typer.Argument(..., help="Name of the model to use from Ollama.")):
     """

lesa/core/conversations.py

@@ -49,22 +49,6 @@ def _setup_conversation_pipeline(self, document_model: str):
         """Set up the conversation pipeline connections."""
 
         pass
-
-    def embed_documents(self, documents_path: str = None):
-        """
-        Embed documents from a specified path.
-        
-        :param documents_path: Path to documents. If None, uses the default output_dir from initialization.
-        """
-        pass
-
-    def embed_document(self, document_path: str = None):
-        """
-        Embed documents from a specified path.
-        
-        :param documents_path: Path to documents. If None, uses the default output_dir from initialization.
-        """
-        pass
 
     def _chat(self, chain, system_prompt: Optional[str] = None):
         """
@@ -157,4 +141,17 @@ def start_conversation(self):
         """
         Start a conversation with the RAG pipeline.
         """
-        self._chat(self.pipe)
+
+        llm = self.ollama_manager.serve_llm()
+        prompt = ChatPromptTemplate([
+            ("system", """Answer any use questions based solely on the context below:
+
+<context>
+{context}
+</context>"""),
+            ("human", """{input}"""),
+        ])
+
+        combine_docs_chain = create_stuff_documents_chain(llm=llm, prompt=prompt)
+        qa_chain = create_retrieval_chain(retriever=self.embeddings_manager.vector_store.as_retriever(), combine_docs_chain=combine_docs_chain)
+        return self._chat(qa_chain)

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "lesa"
-version = "0.1.0.4"
+version = "0.1.0.5"
 description = "A CLI tool to converse with any document locally using Ollama."
 authors = ["Shantanu Wable <shantanuwable2003@gmail.com>", "Omkar Wagholikar <omkarrwagholikar@gmail.com>"]
 repository = "https://github.com/shxntanu/lesa"