Add financial report analysis

fingpt/FinGPT_FinancialReportAnalysis/README.md

@@ -0,0 +1,48 @@
+# Financial Report Analysis Project
+
+## Overview
+
+This project provides tools for analyzing financial reports, specifically annual reports (10-K), using advanced language models such as GPT-4 or other locally deployed Large Language Models (LLM). It's designed to help users generate comprehensive analysis reports in PDF format, offering insights into a company's financial health and performance over the fiscal year.
+
+## Features
+
+- **PDF Report Generation**: Automatically generate detailed analysis reports in PDF format for annual financial statements.
+- **GPT-4 and LLM Support**: Utilize the power of GPT-4 or any locally deployed LLM for deep and insightful analysis.
+- **RAG Support**: The ability to utilize the power of RAG for question-answering and summarization tasks.
+- **Customizable Analysis**: Users can modify the analysis scope by choosing different company symbols and models.
+- **Easy to Use**: Designed with simplicity in mind, simply run all cells in the provided notebook to get your report.
+
+## Requirements
+
+Before starting, ensure you have the following installed:
+- Python 3.11 or later
+- Jupyter Notebook
+- Necessary Python packages (pandas, matplotlib, openai, etc.)
+
+## Getting Started
+
+To begin analyzing financial reports:
+
+0. **(optional) Clone the Repository**:
+   If you want to run the analysis with the locally deployed models, please download Ollama and have it running: https://ollama.com/download.
+   Also, download the model you want to use in the list of available models: https://ollama.com/library with command:
+   ```bash
+    ollama run <model_name>
+    ```
+
+1. **Open the Notebook**:
+   Launch Jupyter Notebook and open the `reportanalysis.ipynb` file:
+   ```
+   jupyter notebook reportanalysis.ipynb
+   ```
+   All the necessary libraries and dependencies are already imported in the notebook.
+
+2. **Configure the Notebook**:
+   Modify the `company symbol` and `models` variables within the notebook to suit the analysis you wish to perform.
+
+3. **Run the Analysis**:
+   Execute all cells in the notebook to generate your financial report analysis in PDF format.
+
+## Contributing
+
+We welcome contributions and suggestions! Please open an issue or submit a pull request with your improvements.

fingpt/FinGPT_FinancialReportAnalysis/utils/__init__.py

@@ -0,0 +1,2 @@
+from utils.earning_calls import get_earnings_transcript, extract_speakers
+from utils.rag import Raptor

fingpt/FinGPT_FinancialReportAnalysis/utils/earning_calls.py

@@ -0,0 +1,69 @@
+from tenacity import retry, stop_after_attempt, wait_random_exponential
+import requests
+import json
+from datetime import datetime
+import re
+from typing import List
+
+
+def correct_date(yr, dt):
+    """Some transcripts have incorrect date, correcting it
+
+    Args:
+        yr (int): actual
+        dt (datetime): given date
+
+    Returns:
+        datetime: corrected date
+    """
+    dt = datetime.strptime(dt, "%Y-%m-%d %H:%M:%S")
+    if dt.year != yr:
+        dt = dt.replace(year=yr)
+    return dt.strftime("%Y-%m-%d %H:%M:%S")
+
+
+def extract_speakers(cont: str) -> List[str]:
+    """Extract the list of speakers
+
+    Args:
+        cont (str): transcript content
+
+    Returns:
+        List[str]: list of speakers
+    """
+    pattern = re.compile(r"\n(.*?):")
+    matches = pattern.findall(cont)
+
+    return list(set(matches))
+
+
+@retry(wait=wait_random_exponential(min=1, max=5), stop=stop_after_attempt(2))
+def get_earnings_transcript(quarter: str, ticker: str, year: int):
+    """Get the earnings transcripts
+
+    Args:
+        quarter (str)
+        ticker (str)
+        year (int)
+    """
+    response = requests.get(
+        f"https://discountingcashflows.com/api/transcript/{ticker}/{quarter}/{year}/",
+        auth=("user", "pass"),
+    )
+
+    resp_text = json.loads(response.text)
+    # speakers_list = extract_speakers(resp_text[0]["content"])
+    corrected_date = correct_date(resp_text[0]["year"], resp_text[0]["date"])
+    resp_text[0]["date"] = corrected_date
+    return resp_text[0]
+
+
+
+# from utils import get_earnings_transcript
+
+# quarter = "Q4"
+# ticker = "AAPL"
+# year = 2023
+# resp_dict, speakers_list = get_earnings_transcript(
+#             quarter, ticker, year
+#         )