Update README.md
Browse files
README.md
CHANGED
|
@@ -9,6 +9,64 @@ app_file: app.py
|
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
# 7/9 - evaluate new GPT models
|
| 13 |
|
| 14 |
GPT-4o Documentation: https://cookbook.openai.com/examples/gpt4o/introduction_to_gpt4o
|
|
|
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
---
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
| 🧩 **Category** | **Info** |
|
| 15 |
+
|---------------------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------|
|
| 16 |
+
| 🛠️ **Libraries & Modules Used** | **Web & UI**: streamlit, streamlit.components.v1 |
|
| 17 |
+
| | **AI & API Integration**: openai, gradio_client |
|
| 18 |
+
| | **File Handling**: base64, os, glob, zipfile, textract, PyPDF2 |
|
| 19 |
+
| | **Image & Video Processing**: cv2, moviepy, PIL |
|
| 20 |
+
| | **Text Processing**: re, BeautifulSoup, pandas |
|
| 21 |
+
| | **Time & Date**: datetime, pytz |
|
| 22 |
+
| | **Utilities & Concurrency**: concurrent.futures, ThreadPoolExecutor, tqdm |
|
| 23 |
+
| | **Speech & Audio**: audio_recorder_streamlit |
|
| 24 |
+
| ⚙️ **App Configuration** | **Site Name**: Scholarly-Article-Document-Search-With-Memory |
|
| 25 |
+
| | **Page Title/Icon**: 🔬🧠ScienceBrain.AI, custom icon file (icons.ico) |
|
| 26 |
+
| | **Sidebar**: Save session checkbox (should_save) |
|
| 27 |
+
| 🗣️ **Core Functionalities** | **Text Interaction**: |
|
| 28 |
+
| | - Chat-based prompts using GPT-4o. |
|
| 29 |
+
| | - Saves conversations as Markdown files (md). |
|
| 30 |
+
| | - Speech Synthesis (HTML5) |
|
| 31 |
+
| | - Embedded JavaScript function (SpeechSynthesis) to read aloud content. |
|
| 32 |
+
| | **Image Interaction**: |
|
| 33 |
+
| | - Upload image, base64 encode, and analyze via GPT-4o. |
|
| 34 |
+
| | - Results stored as Markdown (md) with filenames including prompts and timestamps. |
|
| 35 |
+
| | **Audio Interaction**: |
|
| 36 |
+
| | - Upload or record audio. |
|
| 37 |
+
| | - Audio transcribed using Whisper, summarized or analyzed via GPT-4o, and responses stored as Markdown files. |
|
| 38 |
+
| | **Video Interaction**: |
|
| 39 |
+
| | - Extracts frames and audio from videos. |
|
| 40 |
+
| | - Transcribes audio track, summarizes via GPT-4o. |
|
| 41 |
+
| | - Markdown files created with summarized content. |
|
| 42 |
+
| 📚 **Advanced Document Handling** | **Vector Stores & PDF Galleries**: |
|
| 43 |
+
| | - Allows upload of multiple PDF files, generating quizzes, summaries, or key facts. |
|
| 44 |
+
| | - Vector stores created and managed for RAG querying. |
|
| 45 |
+
| | - Arxiv scholarly search integration through Hugging Face Gradio API (awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern). |
|
| 46 |
+
| | **RAG (Retrieval-Augmented Generation)**: |
|
| 47 |
+
| | - Performs semantic search on uploaded PDFs. |
|
| 48 |
+
| | - Evaluates retrieval performance metrics (recall@k, mrr, map). |
|
| 49 |
+
| 🗃️ **File Management** | **Sidebar File Manager**: |
|
| 50 |
+
| | - Filter files (.md, .pdf, .png, etc.). |
|
| 51 |
+
| | - Operations: View, Edit, Download, Run, Delete individual or bulk actions. |
|
| 52 |
+
| | - Zip download functionality for filtered files. |
|
| 53 |
+
| 🛠️ **Helper Functions & Utilities** | **Filename Generation (generate_filename)**: |
|
| 54 |
+
| | - Date/time, prompt sanitized for filesystem. |
|
| 55 |
+
| | **File Saving (create_and_save_file)**: |
|
| 56 |
+
| | - Conditional file writing based on checkbox option. |
|
| 57 |
+
| | **Concurrent PDF uploads**: |
|
| 58 |
+
| | - Managed via ThreadPoolExecutor with progress tracking. |
|
| 59 |
+
| 🎨 **UI Enhancements** | - Chat-style user-assistant interactions. |
|
| 60 |
+
| | - Streamlit's columns for organized interface and clean button prompts. |
|
| 61 |
+
| | - Dynamic HTML and JavaScript embeds for rich user interaction. |
|
| 62 |
+
| 🌐 **Environmental Variables** | - **Required variables**: |
|
| 63 |
+
| | - API_KEY |
|
| 64 |
+
| | - HF_KEY (Hugging Face) |
|
| 65 |
+
| | - OPENAI_API_KEY |
|
| 66 |
+
| | - OPENAI_ORG_ID |
|
| 67 |
+
| 📝 **Note** | The app robustly integrates multimodal capabilities (text, audio, video, images) and leverages OpenAI and Hugging Face resources extensively to provide a rich, interactive AI-driven document management and analysis experience. |
|
| 68 |
+
|
| 69 |
+
|
| 70 |
# 7/9 - evaluate new GPT models
|
| 71 |
|
| 72 |
GPT-4o Documentation: https://cookbook.openai.com/examples/gpt4o/introduction_to_gpt4o
|