cognicore 0.1.9

A Python toolbox for easy interaction with various LLMs and Vision models

CogniCore 🚀

A Python package that provides a unified, easy-to-use interface for working with various Language Models (LLMs) and Vision Models from multiple providers. 🎯 It focuses on leveraging the generous free tiers offered by AI platforms.

Features

• Text generation with multiple LLM providers support
• Image analysis and description capabilities
• Support for models like Llama, Groq, and Google's Gemini
• Streaming responses
• Tool integration support
• JSON output formatting
• Customizable system prompts

Installation 💻

uv pip install cognicore

Configuration ⚙️

Before using the library, you need to configure your API keys in a .env file:

GROQ_API_KEY=your_groq_key
GITHUB_TOKEN=your_github_token
GOOGLE_API_KEY=your_google_key
SAMBANOVA_API_KEY=your_sambanova_key
CEREBRAS_API_KEY=your_cerebras_key

Quick Start

Text Generation

from cognicore import LanguageModel

# Initialize a session with your preferred model
session = LanguageModel(
    model_name="gemini-2.0-flash",
    provider="google",
    temperature=0.7
)

# Generate a response
response = session.answer("What is the capital of France?")
print(response)

Image Analysis

from cognicore import ImageAnalyzerAgent

analyzer = ImageAnalyzerAgent()
description = analyzer.describe(
    "path/to/image.jpg",
    prompt="Describe the image",
    vllm_provider="groq",
    vllm_name="llama-3.2-90b-vision-preview"
)
print(description)

Usage 🎮

Text Models 📚

from cognicore import LanguageModel

# Initialize a session with your preferred model
session = LanguageModel(
    model_name="llama-3-70b",
    provider="groq",
    temperature=0.7,
    top_k=45,
    top_p=0.95
)

# Simple text generation
response = session.answer("What is the capital of France?")

# JSON-formatted response with Pydantic validation
from pydantic import BaseModel

class LocationInfo(BaseModel):
    city: str
    country: str
    description: str

response = session.answer(
    "What is the capital of France?",
    json_formatting=True,
    pydantic_object=LocationInfo
)

# Using custom tools
tools = [
    {
        "name": "weather",
        "description": "Get current weather",
        "function": get_weather
    }
]
response, tool_calls = session.answer(
    "What's the weather in Paris?",
    tool_list=tools
)

# Streaming responses
for chunk in session.answer(
    "Tell me a long story.",
    stream=True
):
    print(chunk, end="", flush=True)

Vision Models 👁️

from cognicore import ImageAnalyzerAgent

# Initialize the agent
analyzer = ImageAnalyzerAgent()

# Analyze an image
description = analyzer.describe(
    image_path="path/to/image.jpg",
    prompt="Describe this image in detail",
    vllm_provider="groq"
)
print(description)

Available Models 📊

Note: This list is not exhaustive. The library supports any new model ID released by these providers - you just need to get the correct model ID from your provider's documentation.

Text Models

Provider	Model	LLM Provider ID	Model ID	Price	Rate Limit (per min)	Context Window	Speed
Google	Gemini Pro Exp	google	gemini-2.0-pro-exp-02-05	Free	60	32,768	Ultra Fast
Google	Gemini Flash	google	gemini-2.0-flash	Free	60	32,768	Ultra Fast
Google	Gemini Flash Thinking	google	gemini-2.0-flash-thinking-exp-01-21	Free	60	32,768	Ultra Fast
Google	Gemini Flash Lite	google	gemini-2.0-flash-lite-preview-02-05	Free	60	32,768	Ultra Fast
GitHub	O3 Mini	github	o3-mini	Free	50	8,192	Fast
GitHub	GPT-4o	github	gpt-4o	Free	50	8,192	Fast
GitHub	GPT-4o Mini	github	gpt-4o-mini	Free	50	8,192	Fast
GitHub	O1 Mini	github	o1-mini	Free	50	8,192	Fast
GitHub	O1 Preview	github	o1-preview	Free	50	8,192	Fast
GitHub	Meta Llama 3.1 405B	github	meta-Llama-3.1-405B-Instruct	Free	50	8,192	Fast
GitHub	DeepSeek R1	github	DeepSeek-R1	Free	50	8,192	Fast
Groq	DeepSeek R1 Distill Llama 70B	groq	deepseek-r1-distill-llama-70b	Free	100	131,072	Ultra Fast
Groq	Llama 3.3 70B Versatile	groq	llama-3.3-70b-versatile	Free	100	131,072	Ultra Fast
Groq	Llama 3.1 8B Instant	groq	llama-3.1-8b-instant	Free	100	131,072	Ultra Fast
Groq	Llama 3.2 3B Preview	groq	llama-3.2-3b-preview	Free	100	131,072	Ultra Fast
SambaNova	Llama3 405B	sambanova	llama3-405b	Free	60	8,000	Fast

Vision Models

Provider	Model	Vision Provider ID	Model ID	Price	Rate Limit (per min)	Speed
Google	Gemini Vision Exp	gemini	gemini-exp-1206	Free	60	Ultra Fast
Google	Gemini Vision Flash	gemini	gemini-2.0-flash	Free	60	Ultra Fast
GitHub	GPT-4o Vision	github	gpt-4o	Free	50	Fast
GitHub	GPT-4o Mini Vision	github	gpt-4o-mini	Free	50	Fast

Usage Example with Provider ID and Model ID

from cognicore import LanguageModel

# Initialize a session with specific provider and model IDs
session = LanguageModel(
    model_name="llama-3.3-70b-versatile",  # Model ID from the table above
    provider="groq",                        # Provider ID from the table above
    temperature=0.7
)

Requirements

• Python 3.8 or higher
• Required dependencies will be automatically installed

Key Features ⭐

• Simple and intuitive session-based interface
• Support for both vision and text models
• Simple configuration with .env file
• Automatic context management
• Tool support for compatible models
• JSON output formatting with Pydantic validation
• Response streaming support
• Smart caching system
• CPU and GPU support

Contributing 🤝

Contributions are welcome! Feel free to:

1. Fork the project
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Open a Pull Request

License 📄

This project is licensed under the MIT License. See the LICENSE file for details.

Flexible Configuration ⚡

You can initialize both LanguageModel and ImageAnalyzerAgent in three ways:

1. Manual arguments (classic Python style):

   from cognicore import LanguageModel
   llm = LanguageModel(
       model_name="llama-3.3-70b-versatile",
       provider="groq",
       temperature=0.7,
       max_tokens=1024,
   )
   

2. With a configuration dictionary (useful for programmatic config or dynamic settings):

   config = {
       'model_name': 'llama-3.3-70b-versatile',
       'llm_provider': 'groq',
       'temperature': 0.7,
       'max_tokens': 1024,
   }
   llm = LanguageModel(config=config)
   

3. With a YAML config file path (for reproducibility, sharing, and easy experiment management):

   llm = LanguageModel(config="exemple_config.yaml")
   # The YAML file should contain keys like model_name, llm_provider, temperature, etc.
   

The same logic applies to ImageAnalyzerAgent:

analyzer = ImageAnalyzerAgent(config="exemple_config.yaml")

Why is this useful?

• You can easily switch between experiments by changing a config file, not your code.
• Share your settings with collaborators or for reproducibility.
• Centralize all your model and generation parameters in one place.
• Use the same config for both text and vision models.

Multi-Image Support for Vision Models 🖼️🖼️

For some providers (notably Gemini and Groq), you can pass either a single image path or a list of image paths to the describe method:

# Single image
result = analyzer.describe("path/to/image1.jpg", prompt="Describe this image", vllm_provider="gemini")

# Multiple images (Gemini or Groq only)
result = analyzer.describe([
    "path/to/image1.jpg",
    "path/to/image2.jpg"
], prompt="Describe both images", vllm_provider="groq")

Note: Passing multiple images is only supported for providers that support it (currently Gemini and Groq). For other providers, only a single image path (str) is accepted.

Text Classification Utility: TextClassifier

TextClassifier est une classe utilitaire permettant de classifier un texte parmi une liste de classes définies (index, nom, description). Elle hérite de LanguageModel et repose donc sur la même interface flexible (arguments manuels, dictionnaire de config, ou chemin de config YAML).

• Héritage : TextClassifier hérite de LanguageModel pour profiter de toute la logique d'appel LLM multi-provider.
• Utilisation : Fournissez un dictionnaire de classes (ou configurez-le dans le YAML), et utilisez la méthode .classify() pour obtenir l'index ou le nom de la classe prédite.
• Prompts : Les prompts utilisés pour la classification sont stockés dans le dossier prompts.
• Paramètres : Les paramètres spécifiques à la classification sont à placer dans la section text_classifier_utils.py de la config (voir exemple ci-dessous).

Exemple d'utilisation

from cognicore.text_classifier_utils import TextClassifier

# Utilisation avec un fichier de config YAML
classifier = TextClassifier(config="exemple_config.yaml")
texte = "Je cherche un emploi à Paris."
print("Index de classe :", classifier.classify(texte))
print("Nom de classe :", classifier.classify(texte, return_class_name=True))

Exemple de section config (extrait de exemple_config.yaml)

# Paramètres pour text_classifier_utils.py
classification_labels_dict: {
  0: {"class_name": "question", "description": "A general question about any topic."},
  1: {"class_name": "job_search", "description": "A request related to job searching or job offers."},
  2: {"class_name": "internet_search", "description": "A request to search for information on the internet."}
}
classifier_system_prompt: "You are an agent in charge of classifying user's queries into different categories of tasks."
query_classification_model: "meta-llama/llama-4-scout-17b-16e-instruct"
query_classification_provider: "groq"

• Les paramètres passés explicitement à la classe sont prioritaires sur ceux de la config.
• Les prompts sont à placer dans le dossier prompts.

Image Classification Utility: ImageClassifier

ImageClassifier est une classe utilitaire permettant de classifier une image parmi une liste de classes définies (index, nom, description). Elle hérite de ImageAnalyzerAgent (voir vision_utils.py) et repose donc sur la même interface flexible (arguments manuels, dictionnaire de config, ou chemin de config YAML).

• Héritage : ImageClassifier hérite de ImageAnalyzerAgent pour profiter de toute la logique d'appel vision multi-provider.
• Utilisation : Fournissez un dictionnaire de classes (ou configurez-le dans le YAML), et utilisez la méthode .classify() pour obtenir l'index ou le nom de la classe prédite pour une image.
• Prompts : Les prompts utilisés pour la classification sont stockés dans le dossier prompts.
• Paramètres : Les paramètres spécifiques à la classification d'image sont à placer dans la section image_classifier_utils.py de la config (voir exemple ci-dessous).

Exemple d'utilisation

from cognicore.image_classifier_utils import ImageClassifier

# Utilisation avec un fichier de config YAML
classifier = ImageClassifier(config="exemple_config.yaml")
image_path = "test_data/chat.jpg"
print("Index de classe :", classifier.classify(image_path))
print("Nom de classe :", classifier.classify(image_path, return_class_name=True))

Exemple de section config (extrait de exemple_config.yaml)

# Paramètres pour image_classifier_utils.py
image_classification_labels_dict: {
  0: {"class_name": "animal", "description": "Image contenant un animal."},
  1: {"class_name": "ville", "description": "Image représentant une ville ou un lieu urbain."},
  2: {"class_name": "autre", "description": "Tout autre type d'image."}
}
image_classifier_system_prompt: "You are an agent in charge of classifying images into different categories."
image_classification_model: "gemini-1.5-flash"
image_classification_provider: "gemini"

• Les paramètres passés explicitement à la classe sont prioritaires sur ceux de la config.
• Les prompts sont à placer dans le dossier prompts.

Internet Search Utility: InternetSearcher

InternetSearcher is a utility class designed to leverage LLMs with web-browsing capabilities to answer queries based on up-to-date information from the internet. It currently uses Groq's compound-beta model, which can access the web.

• Functionality: Takes a prompt and returns an answer synthesized from internet search results.
• Configuration: Like other utilities in this toolbox, it can be configured via direct arguments, a dictionary, or a YAML file.
• Streaming: Supports streaming responses for real-time output.

Usage Example

from cognicore import InternetSearcher

# Initialize with default settings (Groq's compound-beta)
searcher = InternetSearcher()

# Or initialize with a config file
# searcher = InternetSearcher(config="exemple_config.yaml")

prompt = "What are the latest advancements in battery technology for electric vehicles in 2024?"

# Get a direct answer
result = searcher.search(prompt)
print(result)

# Stream the answer
print("\n--- Streaming ---")
for chunk in searcher.search(prompt, stream=True):
    print(chunk, end="", flush=True)

Configuration in exemple_config.yaml

To configure the InternetSearcher, you can add the following keys to your YAML file:

# Parameters for InternetSearcher
internet_search_model: "compound-beta"
internet_search_provider: "groq"
# You can also override temperature, max_tokens, etc.