Include Google ManimML examples

README.md

@@ -44,4 +44,10 @@ Create beautiful explanatory animations from simple text prompts using AI and Ma
 Set your Gemini API key as an environment variable:
 ```bash
 export GEMINI_API_KEY="your_api_key_here"
-```
+```
+
+Install dependencies (includes **manim-ml** for ML visualizations):
+```bash
+pip install -r requirements.txt
+```
+You can explore additional examples in [Google's ManimML code samples](https://github.com/google/manim-ml/tree/main/examples).

demo.py

@@ -15,7 +15,7 @@
 
 Запуск:
 ```bash
-pip install --upgrade gradio google-genai manim_video_generator
+pip install --upgrade gradio google-genai manim_video_generator manim-ml
 export GEMINI_API_KEY="YOUR_KEY"
 python gradio_manim_gemini_app.py
 ```

manim_video_generator/gemini_client.py

@@ -27,6 +27,8 @@ class GeminiClient:
         
         # Системный промпт с инструкциями
         system_prompt = """You are a Manim code generator. Your ONLY task is to generate executable Python code using the Manim library.
+You can also use the `manim_ml` package (ManimML) for ready-made machine learning scenes and helpers.
+Feel free to reuse examples from Google's official ManimML code samples to illustrate complex ML visuals.
 
 CRITICAL RULES:
 - You MUST respond with ONLY Python code, nothing else

packages.txt

@@ -25,4 +25,5 @@ libopenjp2-7-dev
 libtiff5-dev
 gir1.2-glib-2.0
 libgirepository-1.0-1
-libgtk-3-dev 
+libgtk-3-dev
+graphviz

prompts.py

@@ -9,6 +9,8 @@ Answer is a normal text, not a code. If you have any questions, ask user for cla
 
 SYSTEM_PROMPT_CODEGEN = """
 Now you are a Manim code generator. Your ONLY task is to generate executable Python code using the Manim library.
+You also have access to the `manim_ml` package (ManimML) which provides ready-made scenes and helpers for machine learning visualizations such as neural network diagrams.
+Feel free to reuse snippets from Google's official ManimML code samples if that helps illustrate the request.
 User will provide you a video idea that he has discussed with scenario manager and you will need to generate a Manim code that will execute the user request.
 Format code inside ```python and ``` tags. Answer only code, no other text. Generate your code in one file. P.S. Do not use latex for text.
 """

requirements.txt

@@ -5,4 +5,5 @@ moviepy==1.0.3
 loguru
 numpy
 matplotlib
-Pillow
+Pillow
+manim-ml