[PR #3] [MERGED] Image to Text Module, WhatsApp handler adapted to LangGraph & Docker support #9

New issue

Closed

opened 2026-03-03 16:41:18 +03:00 by kerem · 0 comments

kerem commented

2026-03-03 16:41:18 +03:00

Owner

📋 Pull Request Information

Original PR: https://github.com/neural-maze/ava-whatsapp-agent-course/pull/3
Author: @jesuscopado
Created: 1/3/2025
Status: ✅ Merged
Merged: 1/3/2025
Merged by: @jesuscopado

Base: main ← Head: jesuscopado/image-to-text

📝 Commits (10+)

080f06b Improved character card prompt, added communication style section to focus on how the messages should be and gave context about physical body in the scenario section
6e73587 Image to text module working on chainlit interface using llama 3.2 via groq
3c0b111 Updated elevenlabs model to use newer 2.5 flash model
ebcd764 Ruff formatting
49845a9 Added Docker support with Dockerfile and docker-compose.yml; update FastAPI mount path
5cf072b Fixed Pylance formatting issue, Literal cannot be a variable
7635383 Updated session management using chainlit session_id as thread_id in langgraph. Refactored chainlit_app to use global instances for modules.
5ec85b1 Refactored WhatsApp response adapting to new langgraph agent
b1604c8 Improved the image generation logic and prompt to utilize the last five messages for context
2aa53c1 Improved router prompt with more explicit rules and examples

📊 Changes

16 files changed (+403 additions, -170 deletions)

View changed files

📝 .gitignore (+1 -1)
➕ Dockerfile (+30 -0)
➕ docker-compose.yml (+11 -0)
📝 src/.chainlit/config.toml (+1 -1)
📝 src/ai_companion/core/exceptions.py (+6 -0)
📝 src/ai_companion/core/prompts.py (+75 -50)
📝 src/ai_companion/graph/utils/edges.py (+1 -1)
📝 src/ai_companion/graph/utils/helpers.py (+5 -0)
📝 src/ai_companion/graph/utils/nodes.py (+11 -5)
📝 src/ai_companion/modules/image/__init__.py (+4 -0)
📝 src/ai_companion/modules/image/image_to_text.py (+105 -1)
📝 src/ai_companion/modules/image/text_to_image.py (+18 -15)
📝 src/ai_companion/settings.py (+2 -1)
📝 src/interfaces/chainlit_app.py (+43 -18)
📝 src/interfaces/fastapi_app.py (+1 -1)
📝 src/interfaces/whatsapp_response.py (+89 -76)

📄 Description

No description provided

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/neural-maze/ava-whatsapp-agent-course/pull/3 **Author:** [@jesuscopado](https://github.com/jesuscopado) **Created:** 1/3/2025 **Status:** ✅ Merged **Merged:** 1/3/2025 **Merged by:** [@jesuscopado](https://github.com/jesuscopado) **Base:** `main` ← **Head:** `jesuscopado/image-to-text` --- ### 📝 Commits (10+) - [`080f06b`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/080f06b90ac931f44ee8e31179ccee422a3996d4) Improved character card prompt, added communication style section to focus on how the messages should be and gave context about physical body in the scenario section - [`6e73587`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/6e7358728def892bae4518e2d380957fbcc0e344) Image to text module working on chainlit interface using llama 3.2 via groq - [`3c0b111`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/3c0b111c3a88c4fa6531d8f97ec56f4d63bcf0e0) Updated elevenlabs model to use newer 2.5 flash model - [`ebcd764`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/ebcd7648d70fba92529902ffd5943cd1bb71bd8e) Ruff formatting - [`49845a9`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/49845a918a6e0550bc131f73d47119661dc6a7ed) Added Docker support with Dockerfile and docker-compose.yml; update FastAPI mount path - [`5cf072b`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/5cf072baa5fba8640db06637c1ff5b05f00d3f6c) Fixed Pylance formatting issue, Literal cannot be a variable - [`7635383`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/763538321d73e5625388c677b089ebfe701483af) Updated session management using chainlit session_id as thread_id in langgraph. Refactored chainlit_app to use global instances for modules. - [`5ec85b1`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/5ec85b105b1b97eea9b7325d2d9e87d02e7d8b85) Refactored WhatsApp response adapting to new langgraph agent - [`b1604c8`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/b1604c85a967b3af9bae41bd4318e80917fb8405) Improved the image generation logic and prompt to utilize the last five messages for context - [`2aa53c1`](https://github.com/neural-maze/ava-whatsapp-agent-course/commit/2aa53c1471f615c3d72d8a63318c15f57dacb028) Improved router prompt with more explicit rules and examples ### 📊 Changes **16 files changed** (+403 additions, -170 deletions) <details> <summary>View changed files</summary> 📝 `.gitignore` (+1 -1) ➕ `Dockerfile` (+30 -0) ➕ `docker-compose.yml` (+11 -0) 📝 `src/.chainlit/config.toml` (+1 -1) 📝 `src/ai_companion/core/exceptions.py` (+6 -0) 📝 `src/ai_companion/core/prompts.py` (+75 -50) 📝 `src/ai_companion/graph/utils/edges.py` (+1 -1) 📝 `src/ai_companion/graph/utils/helpers.py` (+5 -0) 📝 `src/ai_companion/graph/utils/nodes.py` (+11 -5) 📝 `src/ai_companion/modules/image/__init__.py` (+4 -0) 📝 `src/ai_companion/modules/image/image_to_text.py` (+105 -1) 📝 `src/ai_companion/modules/image/text_to_image.py` (+18 -15) 📝 `src/ai_companion/settings.py` (+2 -1) 📝 `src/interfaces/chainlit_app.py` (+43 -18) 📝 `src/interfaces/fastapi_app.py` (+1 -1) 📝 `src/interfaces/whatsapp_response.py` (+89 -76) </details> ### 📄 Description _No description provided_ --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>