MCP-powered Financial Analyst using Langgraph #126
Conversation
added readme.md
added .env file
WalkthroughA new financial analysis workflow project was introduced, featuring environment setup, documentation, and a modular codebase. The workflow parses user queries about stock analysis, generates corresponding Python code using Azure OpenAI, and enables code execution and visualization. A server interface provides tools for query analysis, code saving, and code execution. Dependency management and setup instructions are included. Changes
Sequence Diagram(s)sequenceDiagram
participant User
participant Server
participant Workflow
participant AzureOpenAI
User->>Server: analyze_stock(query)
Server->>Workflow: run_financial_analysis(query)
Workflow->>AzureOpenAI: Parse query (QueryParser node)
AzureOpenAI-->>Workflow: Parsed fields (symbol, timeframe, action)
Workflow->>AzureOpenAI: Generate Python code (CodeWriter node)
AzureOpenAI-->>Workflow: Python code
Workflow-->>Server: Return generated code
Server-->>User: Show generated code
User->>Server: save_code(code)
Server->>Server: Write code to stock_analysis.py
Server-->>User: Return success or error message
User->>Server: run_code_and_show_plot()
Server->>Server: Execute stock_analysis.py
Server-->>User: Display plot or result
Poem
✨ Finishing Touches
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 10
♻️ Duplicate comments (3)
financial-analyst-deepseek/financial-analyst-langgraph/building-financial-analyst-langgraph.ipynb (3)
9-21: 🛠️ Refactor suggestionClean up imports in the notebook.
Similar to the Python module, this notebook has several unused imports that should be removed for clarity.
130-130:⚠️ Potential issueFix typo: "recieved" → "received"
147-147:⚠️ Potential issueFix critical syntax error in code_result function.
Same issue as in the Python module - incorrect dictionary access.
def code_result(state: StockAnalysisState): - ans=StockAnalysisState["generated_code"] + ans = state["generated_code"] return {"execution_result": ans}
🧹 Nitpick comments (7)
financial-analyst-deepseek/financial-analyst-langgraph/README.md (2)
15-24: Fix markdown formatting issues.The static analysis tools identified several formatting issues that should be addressed for better readability.
Apply these formatting improvements:
-**Fill Your Environment Variables** +## Fill Your Environment Variables A `.env` file is already included in the project. -Open the file and fill in your actual API keys: +Open the file and fill in your actual API keys. + +Copy the example environment file and configure your credentials: +```bash +cp .env.example .env +# Edit .env with your actual Azure OpenAI credentials +``` -**Install Dependencies** +## Install Dependencies Ensure you have Python 3.12 or later installed. -``` +```bash pip install -r requirements.txt<details> <summary>🧰 Tools</summary> <details> <summary>🪛 markdownlint-cli2 (0.17.2)</summary> 15-15: Emphasis used instead of a heading null (MD036, no-emphasis-as-heading) --- 24-24: Emphasis used instead of a heading null (MD036, no-emphasis-as-heading) </details> </details> --- `59-59`: **Add missing comma for better readability.** ```diff -In Cursor MCP settings make sure to toggle the button to connect the server to the host. Done! Your server is now up and running. +In Cursor MCP settings, make sure to toggle the button to connect the server to the host. Done! Your server is now up and running.🧰 Tools
🪛 LanguageTool
[uncategorized] ~59-~59: A comma might be missing here.
Context: ...ted in the MCP settings. In Cursor MCP settings make sure to toggle the button to conne...(AI_EN_LECTOR_MISSING_PUNCTUATION_COMMA)
financial-analyst-deepseek/financial-analyst-langgraph/server.py (1)
27-31: Enhance error handling in analyze_stock function.The current error handling is minimal and doesn't provide enough context for debugging issues.
try: result = run_financial_analysis(query) + if not result: + return "Error: No analysis result generated" return result except Exception as e: + import traceback + error_details = traceback.format_exc() + print(f"Analysis error: {error_details}") # Log for debugging return f"Error: {e}"financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (1)
56-61: Consider extracting the prompt to improve readability.The prompt string is quite long and contains formatting instructions. Consider moving it to a constant or separate function for better maintainability.
+CODE_WRITER_PROMPT = """You are a Senior Python Developer. Generate code to {action} the stock data. +Stock: {symbol} +Timeframe: {timeframe} + +Use yfinance, pandas, and matplotlib libraries. Output should be a clean, executable .py Python script for stock visualization without explanations or AI-generated messages—just the direct script content, without ''' or any code blockers.""" + def code_writer_node(state: StockAnalysisState): parsed = state["parsed_output"] if isinstance(parsed, dict): raise NodeInterrupt("received wrong type") - fprompt = """You are a Senior Python Developer. Generate code to {action} the stock data. - Stock: {symbol} - Timeframe: {timeframe} - - Use yfinance, pandas, and matplotlib libraries. Output should be a clean, executable .py Python script for stock visualization without explanations or AI-generated messages—just the direct script content. without ''' or any code blockers - """ + action=parsed.result.action symbol=parsed.result.symbol time=parsed.result.timeframe - ffprompt=fprompt.format(action=action,symbol=symbol,timeframe=time) + ffprompt=CODE_WRITER_PROMPT.format(action=action,symbol=symbol,timeframe=time) code = llm.invoke(ffprompt) return {"generated_code": code}financial-analyst-deepseek/financial-analyst-langgraph/building-financial-analyst-langgraph.ipynb (3)
24-30: Remove empty code cell.This empty cell serves no purpose and should be removed to keep the notebook clean.
135-135: Improve prompt formatting.The prompt has awkward phrasing "without ''' or any code blockers". Consider rephrasing to be clearer.
- Use yfinance, pandas, and matplotlib libraries. Output should be a clean, executable .py Python script for stock visualization without explanations or AI-generated messages—just the direct script content.without ''' or any code blockers + Use yfinance, pandas, and matplotlib libraries. Output should be a clean, executable Python script for stock visualization without explanations or AI-generated messages—just the direct script content without code block markers.
247-250: Remove empty cells at the end of notebook.These empty cells at the end serve no purpose and should be removed.
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
financial-analyst-deepseek/financial-analyst-langgraph/graph.pngis excluded by!**/*.png
📒 Files selected for processing (6)
financial-analyst-deepseek/financial-analyst-langgraph/.env(1 hunks)financial-analyst-deepseek/financial-analyst-langgraph/README.md(1 hunks)financial-analyst-deepseek/financial-analyst-langgraph/building-financial-analyst-langgraph.ipynb(1 hunks)financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py(1 hunks)financial-analyst-deepseek/financial-analyst-langgraph/requirements.txt(1 hunks)financial-analyst-deepseek/financial-analyst-langgraph/server.py(1 hunks)
🧰 Additional context used
🪛 LanguageTool
financial-analyst-deepseek/financial-analyst-langgraph/README.md
[uncategorized] ~59-~59: A comma might be missing here.
Context: ...ted in the MCP settings. In Cursor MCP settings make sure to toggle the button to conne...
(AI_EN_LECTOR_MISSING_PUNCTUATION_COMMA)
🪛 markdownlint-cli2 (0.17.2)
financial-analyst-deepseek/financial-analyst-langgraph/README.md
15-15: Emphasis used instead of a heading
null
(MD036, no-emphasis-as-heading)
24-24: Emphasis used instead of a heading
null
(MD036, no-emphasis-as-heading)
27-27: Fenced code blocks should have a language specified
null
(MD040, fenced-code-language)
🪛 Ruff (0.11.9)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
3-3: langchain_core.runnables.Runnable imported but unused
Remove unused import: langchain_core.runnables.Runnable
(F401)
4-4: typing.Literal imported but unused
Remove unused import: typing.Literal
(F401)
5-5: langchain_core.messages.AIMessage imported but unused
Remove unused import
(F401)
5-5: langchain_core.messages.HumanMessage imported but unused
Remove unused import
(F401)
6-6: ast imported but unused
Remove unused import: ast
(F401)
7-7: IPython.display.Image imported but unused
Remove unused import
(F401)
7-7: IPython.display.display imported but unused
Remove unused import
(F401)
8-8: typing.Dict imported but unused
Remove unused import: typing.Dict
(F401)
13-13: yfinance imported but unused
Remove unused import: yfinance
(F401)
98-98: Found useless attribute access. Either assign it to a variable or remove it.
(B018)
🔇 Additional comments (3)
financial-analyst-deepseek/financial-analyst-langgraph/server.py (1)
2-2:✅ Verification successful
Verify the imported module exists.
The import references
finance_langgraphmodule which is not provided in this review. Ensure this module exists and contains therun_financial_analysisfunction.
🏁 Script executed:
#!/bin/bash # Description: Check if the finance_langgraph module exists and contains the required function # Expected: Find the file and the function definition fd -t f "finance_langgraph.py" if [ $? -eq 0 ]; then echo "=== Checking for run_financial_analysis function ===" rg -A 5 "def run_financial_analysis" else echo "finance_langgraph.py not found in repository" fiLength of output: 1359
finance_langgraph module found
Confirmed thatfinancial-analyst-langgraph/finance_langgraph.pydefines therun_financial_analysisfunction. No further action required.financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (1)
26-37: Well-structured data models!The Pydantic models and TypedDict are well-designed with clear field descriptions.
financial-analyst-deepseek/financial-analyst-langgraph/building-financial-analyst-langgraph.ipynb (1)
225-242: Good demonstration of workflow execution!The notebook effectively demonstrates the end-to-end workflow by executing the generated code and displaying the stock chart.
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Show resolved
Hide resolved
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Show resolved
Hide resolved
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Outdated
Show resolved
Hide resolved
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Outdated
Show resolved
Hide resolved
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Show resolved
Hide resolved
There was a problem hiding this comment.
Actionable comments posted: 0
♻️ Duplicate comments (5)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (5)
1-13: Remove unused imports to improve code clarity.Multiple imports are not used in the code and should be removed to improve maintainability and reduce confusion.
🧰 Tools
🪛 Ruff (0.11.9)
3-3:
langchain_core.runnables.Runnableimported but unusedRemove unused import:
langchain_core.runnables.Runnable(F401)
4-4:
typing.Literalimported but unusedRemove unused import:
typing.Literal(F401)
5-5:
langchain_core.messages.AIMessageimported but unusedRemove unused import
(F401)
5-5:
langchain_core.messages.HumanMessageimported but unusedRemove unused import
(F401)
6-6:
astimported but unusedRemove unused import:
ast(F401)
7-7:
IPython.display.Imageimported but unusedRemove unused import
(F401)
7-7:
IPython.display.displayimported but unusedRemove unused import
(F401)
8-8:
typing.Dictimported but unusedRemove unused import:
typing.Dict(F401)
13-13:
yfinanceimported but unusedRemove unused import:
yfinance(F401)
17-22: Add validation for required environment variables.The code directly accesses Azure OpenAI environment variables without validation, which can cause unclear errors if any are missing.
54-55: Fix typo in error message."recieved" should be "received".
70-73: Fix critical syntax error in code_result function.The function incorrectly accesses the TypedDict class instead of the state parameter.
96-99: Remove dead code and add error handling.Line 98 accesses
result["generated_code"].contentwithout using it, and the function lacks error handling.
🧹 Nitpick comments (2)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (2)
43-45: Clean up prompt formatting.The prompt has trailing whitespace and could be formatted more cleanly.
- prompt = """You are a Stock Data Analyst. Extract stock details from this user query: {query}. - - """ + prompt = """You are a Stock Data Analyst. Extract stock details from this user query: {query}."""
62-65: Improve variable naming consistency.Variable names could be more descriptive and consistent to enhance code readability.
- action=parsed.result.action - symbol=parsed.result.symbol - time=parsed.result.timeframe - ffprompt=fprompt.format(action=action,symbol=symbol,timeframe=time) + action = parsed.result.action + symbol = parsed.result.symbol + timeframe = parsed.result.timeframe + final_prompt = fprompt.format(action=action, symbol=symbol, timeframe=timeframe)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.11.9)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
3-3: langchain_core.runnables.Runnable imported but unused
Remove unused import: langchain_core.runnables.Runnable
(F401)
4-4: typing.Literal imported but unused
Remove unused import: typing.Literal
(F401)
5-5: langchain_core.messages.AIMessage imported but unused
Remove unused import
(F401)
5-5: langchain_core.messages.HumanMessage imported but unused
Remove unused import
(F401)
6-6: ast imported but unused
Remove unused import: ast
(F401)
7-7: IPython.display.Image imported but unused
Remove unused import
(F401)
7-7: IPython.display.display imported but unused
Remove unused import
(F401)
8-8: typing.Dict imported but unused
Remove unused import: typing.Dict
(F401)
13-13: yfinance imported but unused
Remove unused import: yfinance
(F401)
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (3)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (3)
1-13: Remove unused imports to improve code clarity.Multiple imports are unused and should be removed as identified by static analysis tools.
🧰 Tools
🪛 Ruff (0.11.9)
3-3:
langchain_core.runnables.Runnableimported but unusedRemove unused import:
langchain_core.runnables.Runnable(F401)
4-4:
typing.Literalimported but unusedRemove unused import:
typing.Literal(F401)
5-5:
langchain_core.messages.AIMessageimported but unusedRemove unused import
(F401)
5-5:
langchain_core.messages.HumanMessageimported but unusedRemove unused import
(F401)
6-6:
astimported but unusedRemove unused import:
ast(F401)
7-7:
IPython.display.Imageimported but unusedRemove unused import
(F401)
7-7:
IPython.display.displayimported but unusedRemove unused import
(F401)
8-8:
typing.Dictimported but unusedRemove unused import:
typing.Dict(F401)
13-13:
yfinanceimported but unusedRemove unused import:
yfinance(F401)
17-22: Add validation for required environment variables.Environment variables are accessed without validation, which can cause unclear errors if any are missing.
54-55: Fix typo in error message.The typo "recieved" should be "received".
🧹 Nitpick comments (2)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (2)
41-50: Improve variable naming and prompt formatting.Consider improving code style for better readability:
def query_parser_node(state: StockAnalysisState): query = state["query"] - prompt = """You are a Stock Data Analyst. Extract stock details from this user query: {query}. - - """ - finalprompt=prompt.format(query=query) - llm_with_struc=llm.with_structured_output(QueryAnalysisOutput) - response = llm_with_struc.invoke(finalprompt) + prompt = """You are a Stock Data Analyst. Extract stock details from this user query: {query}.""" + final_prompt = prompt.format(query=query) + llm_with_struct = llm.with_structured_output(QueryAnalysisOutput) + response = llm_with_struct.invoke(final_prompt) return {"parsed_output": response}
56-67: Improve variable naming consistency.Consider using consistent and descriptive variable names for better code readability:
- fprompt = """You are a Senior Python Developer. Generate code to {action} the stock data. + code_prompt = """You are a Senior Python Developer. Generate code to {action} the stock data. Stock: {symbol} Timeframe: {timeframe} Use yfinance, pandas, and matplotlib libraries. Output should be a clean, executable .py Python script for stock visualization without explanations or AI-generated messages—just the direct script content. without ''' or any code blockers """ - action=parsed.result.action - symbol=parsed.result.symbol - time=parsed.result.timeframe - ffprompt=fprompt.format(action=action,symbol=symbol,timeframe=time) - code = llm.invoke(ffprompt) + action = parsed.result.action + symbol = parsed.result.symbol + timeframe = parsed.result.timeframe + final_prompt = code_prompt.format(action=action, symbol=symbol, timeframe=timeframe) + code = llm.invoke(final_prompt)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.11.9)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
3-3: langchain_core.runnables.Runnable imported but unused
Remove unused import: langchain_core.runnables.Runnable
(F401)
4-4: typing.Literal imported but unused
Remove unused import: typing.Literal
(F401)
5-5: langchain_core.messages.AIMessage imported but unused
Remove unused import
(F401)
5-5: langchain_core.messages.HumanMessage imported but unused
Remove unused import
(F401)
6-6: ast imported but unused
Remove unused import: ast
(F401)
7-7: IPython.display.Image imported but unused
Remove unused import
(F401)
7-7: IPython.display.display imported but unused
Remove unused import
(F401)
8-8: typing.Dict imported but unused
Remove unused import: typing.Dict
(F401)
13-13: yfinance imported but unused
Remove unused import: yfinance
(F401)
🔇 Additional comments (2)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (2)
26-37: LGTM! Well-structured Pydantic models.The model definitions follow proper Pydantic patterns with clear field descriptions and appropriate typing.
76-103: LGTM! Well-structured workflow setup.The LangGraph workflow is properly configured with appropriate node connections and the helper function provides a clean interface.
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Outdated
Show resolved
Hide resolved
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (4)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (4)
1-13: Remove unused imports to improve code clarity.Multiple imports are unused and should be removed as identified by static analysis.
🧰 Tools
🪛 Ruff (0.11.9)
3-3:
langchain_core.runnables.Runnableimported but unusedRemove unused import:
langchain_core.runnables.Runnable(F401)
4-4:
typing.Literalimported but unusedRemove unused import:
typing.Literal(F401)
5-5:
langchain_core.messages.AIMessageimported but unusedRemove unused import
(F401)
5-5:
langchain_core.messages.HumanMessageimported but unusedRemove unused import
(F401)
6-6:
astimported but unusedRemove unused import:
ast(F401)
7-7:
IPython.display.Imageimported but unusedRemove unused import
(F401)
7-7:
IPython.display.displayimported but unusedRemove unused import
(F401)
8-8:
typing.Dictimported but unusedRemove unused import:
typing.Dict(F401)
13-13:
yfinanceimported but unusedRemove unused import:
yfinance(F401)
17-22: Add validation for required environment variables.The code directly accesses Azure OpenAI environment variables without validation, which can cause unclear errors if any are missing.
54-55: Fix typo in error message."recieved" should be "received".
102-105: 🛠️ Refactor suggestionAdd error handling to prevent runtime failures.
The function should handle cases where the workflow fails or expected keys are missing:
def run_financial_analysis(query): - result = workflow.invoke({"query": query}) - - return result["generated_code"].content + try: + result = workflow.invoke({"query": query}) + + if "generated_code" not in result: + raise ValueError("Workflow did not generate code") + + generated_code = result["generated_code"] + if not hasattr(generated_code, 'content'): + raise ValueError("Generated code has no content") + + return generated_code.content + except Exception as e: + raise RuntimeError(f"Financial analysis failed: {str(e)}")
🧹 Nitpick comments (2)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py (2)
46-47: Improve variable naming for better readability.Consider using more descriptive variable names:
- finalprompt=prompt.format(query=query) - llm_with_struc=llm.with_structured_output(QueryAnalysisOutput) + formatted_prompt = prompt.format(query=query) + structured_llm = llm.with_structured_output(QueryAnalysisOutput)
62-65: Improve variable naming and spacing.Variable names could be more descriptive and consistent:
- action=parsed.result.action - symbol=parsed.result.symbol - time=parsed.result.timeframe - ffprompt=fprompt.format(action=action,symbol=symbol,timeframe=time) + action = parsed.result.action + symbol = parsed.result.symbol + timeframe = parsed.result.timeframe + formatted_prompt = fprompt.format(action=action, symbol=symbol, timeframe=timeframe)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.11.9)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
3-3: langchain_core.runnables.Runnable imported but unused
Remove unused import: langchain_core.runnables.Runnable
(F401)
4-4: typing.Literal imported but unused
Remove unused import: typing.Literal
(F401)
5-5: langchain_core.messages.AIMessage imported but unused
Remove unused import
(F401)
5-5: langchain_core.messages.HumanMessage imported but unused
Remove unused import
(F401)
6-6: ast imported but unused
Remove unused import: ast
(F401)
7-7: IPython.display.Image imported but unused
Remove unused import
(F401)
7-7: IPython.display.display imported but unused
Remove unused import
(F401)
8-8: typing.Dict imported but unused
Remove unused import: typing.Dict
(F401)
13-13: yfinance imported but unused
Remove unused import: yfinance
(F401)
financial-analyst-deepseek/financial-analyst-langgraph/finance_langgraph.py
Show resolved
Hide resolved
There was a problem hiding this comment.
Use uv as package manager.
refer this readme: https://github.com/patchy631/ai-engineering-hub/tree/main/mcp-video-rag
There was a problem hiding this comment.
Can you please use UV as a package manager? rest of the things look good
There was a problem hiding this comment.
yes definnetly, i will update
I have implemented the same project — "MCP-powered Financial Analyst using CrewAI" : but this time using LangGraph instead of CrewAI.
Why LangGraph?
*Granular control over the multi-agent workflow.
*It is a graph-based framework, which offers better modularity and flow customization.
*Allows for more explicit step-by-step agent orchestration, which is ideal for complex agentic tasks like financial analysis.
This implementation mirrors the original use case but showcases the potential of LangGraph for building more controllable and structured multi-agent systems.
Summary by CodeRabbit
New Features
Documentation
Chores