frirapi 1.0.0

Trilingual AI API module (English, French, Spanish) synchronized with a Supabase cloud brain and equipped with web fallback search capabilities.

frirapi (Frire AI API)

frirapi (also known as Pancakapi) is a lightweight, trilingual (English, French, Spanish) Python module designed to query, train, and manage a synchronized cloud-based knowledge database powered by Supabase. It also includes built-in fallback web-scraping via DuckDuckGo Search when local data isn't enough.

Originally built for the Frire AI ecosystem, this package allows any Python developer to integrate the AI's core brain directly into their own applications, bots, scripts, or web backends.

---

Features

• Cloud Sync: Instantly fetches and synchronizes structured Q&A data from a Supabase central brain.
• Smart Matching: Uses tokenization, string cleaning, and SequenceMatcher string similarity algorithms to find the best local answer.
• Autonomous Learning: Real-time feedback loop that dynamically adjusts the confidence/reliability scores of answers in the cloud database.
• Web Fallback: Automatically scrapes trusted web sources (Wikipedia, Larousse, Britannica, etc.) if local knowledge confidence is low.
• Trilingual Translation: Automatically handles incoming queries and returns localized answers in English, French, or Spanish using deep-translator.
• Adaptive Response Styles: Choose between courte (short), normale (normal), or detaillee (detailed) outputs.
• Hot-Reload Auto-Updater: Update the package to the latest version at runtime without restarting your script.

---

Installation

Install the package and its requirements using your favorite package manager:

pip install frirapi 
uv pip install firapi
poetry add frirapi@latest

and more package manager...

---

Quick Start Guide

Here is how you can fetch the database and start chatting with the AI core:

import frirapi

# 1. Initialize and load the central brain from the Cloud
print("Loading database...")
brain_database = frirapi.charger_base()

# 2. Define a question and preferred settings
user_question = "What is an Hantavirus?"
response_style = "normale"  # Options: "short", "normal", "detailed"
language = "en"             # Options: "en", "fr", "es"

# 3. Query the AI
response, confidence, exact_match = frirapi.ia_repond(
    question=user_question, 
    BASE_DE_DONNEES=brain_database, 
    style_actuel=response_style, 
    langue_ia=language
)

# 4. Format and print the final result
if response:
    final_output = frirapi.varier_reponse(response, response_style, user_question)
    print(f"AI Response: {final_output}")
    print(f"Confidence Score: {confidence}%")
    print(f"Source/Match: {exact_match}")
else:
    print("The AI does not know the answer.")

---

Database Interaction & Learning Loop

1. Training the AI (Confidence adjustment)

If the AI answered a question using an existing database entry (exact_match), you can supply a feedback loop to let it learn and save the new weight directly to Supabase:

# Assuming the user validated the response:
is_correct = True  # Set to False if the answer was wrong

if exact_match and not exact_match.startswith("[WEB]"):
    frirapi.ia_apprend(exact_match, is_correct, brain_database)

2. Manually adding knowledge to the Cloud

You can contribute directly to the global brain by pushing a new question, answer, and category ("sciences", "maths", or "culture"):

frirapi.ajouter_a_supabase(
    question="What is 2+2?", 
    reponse="4", 
    categorie="maths"
)

---

Hot-Reload Auto-Updater (State)

frirapi comes with an built-in runtime module updater. If you want to make sure your scripts are always using the latest release without manual re-deployments or application restarts, you can call the State.UpdateModule method:

import frirapi

# Update package at runtime using your preferred package environment
# Supported modes: "pip", "uv", "poetry", "conda", "git"
frirapi = frirapi.State.UpdateModule(mode="uv")
# The module is now updated and reloaded into memory!

---

License

This project is licensed under the GNU General Public License v3.0 (GPL-3.0). See the LICENSE file for more details.

Contact & Support

Creator: ZedKa450

Discord: zedka450

E-mail: [e-mail]

GitHub Repository: https://github.com/zedka450/Frire-AI