To generate a Technical report based on inputs of type PDF, DOC, Image, CSV, etc., using Streamlit for UI, unstructured.io for input parsing, LangGraph for Agents to generate content sectionwise and docx to format the final report with images, tables, sections based on user template
- This Repo includes sample inputs/PDFs used, requirements.txt, screenshot of the tool, scripts
- Script uses Azure OpenAI embeddings and LLM(gpt4o)
- Install packages as "python.exe -m pip install -r Py3p11p0_GenAITechReportEditingTool_requirements.txt"
- I'm using unstructured.io for parsing the PDF and Images
- unstructured.io requires tesseract-ocr and poppler, extract, install and mention the installed path in the script
- Modify the paths as required in the scripts
- Run tool script as "python.exe -m streamlit run GenAITechReportGen_Tool_Agent_Pickle.py" for without editing tool and for with editing tool, run "GenAITechReportGen_Tool_Agent_Pickle_TextEditor"