Upload 7 files

README(4).md

@@ -0,0 +1,13 @@
+---
+title: Cassandre
+emoji: 📜
+colorFrom: gray
+colorTo: blue
+sdk: gradio
+sdk_version: 3.50.2
+app_file: app.py
+pinned: false
+license: mit
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app(5).py

@@ -0,0 +1,179 @@
+import transformers
+import re
+from transformers import AutoConfig, AutoTokenizer, AutoModel, AutoModelForCausalLM
+from vllm import LLM, SamplingParams
+import torch
+import gradio as gr
+import json
+import os
+import shutil
+import requests
+import lancedb
+import pandas as pd
+
+# Define the device
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+# Define variables 
+temperature = 0.7
+max_new_tokens = 3000
+top_p = 0.95
+repetition_penalty = 1.2
+
+model_name = "PleIAs/Cassandre-RAG"
+
+# Initialize vLLM
+llm = LLM(model_name, max_model_len=8128)
+
+# Connect to the LanceDB database
+db = lancedb.connect("content19/lancedb_data")
+table = db.open_table("edunat19")
+
+def hybrid_search(text):
+    results = table.search(text, query_type="hybrid").limit(5).to_pandas()
+    
+    # Add a check for duplicate hashes
+    seen_hashes = set()
+    
+    document = []
+    document_html = []
+    for _, row in results.iterrows():
+        hash_id = str(row['hash'])
+        
+        # Skip if we've already seen this hash
+        if hash_id in seen_hashes:
+            continue
+            
+        seen_hashes.add(hash_id)
+        title = row['section']
+        content = row['text']
+
+        document.append(f"**{hash_id}**\n{title}\n{content}")
+        document_html.append(f'<div class="source" id="{hash_id}"><p><b>{hash_id}</b> : {title}<br>{content}</div>')
+
+    document = "\n".join(document)
+    document_html = '<div id="source_listing">' + "".join(document_html) + "</div>"
+    return document, document_html
+
+
+class CassandreChatBot:
+    def __init__(self, system_prompt="Tu es Cassandre, le chatbot de l'Éducation nationale qui donne des réponses sourcées."):
+        self.system_prompt = system_prompt
+
+    def predict(self, user_message):
+        fiches, fiches_html = hybrid_search(user_message)
+        sampling_params = SamplingParams(temperature=temperature, top_p=top_p, max_tokens=max_new_tokens, presence_penalty=repetition_penalty, stop=["#END#"])
+
+        detailed_prompt = f"""### Query ###\n{user_message}\n\n### Source ###\n{fiches}\n\n### Answer ###\n"""
+
+        prompts = [detailed_prompt]
+        outputs = llm.generate(prompts, sampling_params, use_tqdm=False)
+        generated_text = outputs[0].outputs[0].text
+        generated_text = '<h2 style="text-align:center">Réponse</h3>\n<div class="generation">' + format_references(generated_text) + "</div>"
+        fiches_html = '<h2 style="text-align:center">Sources</h3>\n' + fiches_html
+        return generated_text, fiches_html
+
+def format_references(text):
+    ref_start_marker = '<ref text="'
+    ref_end_marker = '</ref>'
+
+    parts = []
+    current_pos = 0
+    ref_number = 1
+
+    while True:
+        start_pos = text.find(ref_start_marker, current_pos)
+        if start_pos == -1:
+            parts.append(text[current_pos:])
+            break
+
+        parts.append(text[current_pos:start_pos])
+
+        end_pos = text.find('">', start_pos)
+        if end_pos == -1:
+            break
+
+        ref_text = text[start_pos + len(ref_start_marker):end_pos].replace('\n', ' ').strip()
+        ref_text_encoded = ref_text.replace("&", "&amp;").replace("<", "&lt;").replace(">", "&gt;")
+
+        ref_end_pos = text.find(ref_end_marker, end_pos)
+        if ref_end_pos == -1:
+            break
+
+        ref_id = text[end_pos + 2:ref_end_pos].strip()
+
+        tooltip_html = f'<span class="tooltip" data-refid="{ref_id}" data-text="{ref_id}: {ref_text_encoded}"><a href="#{ref_id}">[{ref_number}]</a></span>'
+        parts.append(tooltip_html)
+
+        current_pos = ref_end_pos + len(ref_end_marker)
+        ref_number = ref_number + 1
+
+    return ''.join(parts)
+
+# Initialize the CassandreChatBot
+cassandre_bot = CassandreChatBot()
+
+# CSS for styling
+css = """
+.generation {
+    margin-left:2em;
+    margin-right:2em;
+}
+:target {
+    background-color: #CCF3DF;
+  }
+.source {
+    float:left;
+    max-width:17%;
+    margin-left:2%;
+}
+.tooltip {
+    position: relative;
+    cursor: pointer;
+    font-variant-position: super;
+    color: #97999b;
+  }
+  
+  .tooltip:hover::after {
+    content: attr(data-text);
+    position: absolute;
+    left: 0;
+    top: 120%;
+    white-space: pre-wrap;
+    width: 500px;
+    max-width: 500px;
+    z-index: 1;
+    background-color: #f9f9f9;
+    color: #000;
+    border: 1px solid #ddd;
+    border-radius: 5px;
+    padding: 5px;
+    display: block;
+    box-shadow: 0 4px 8px rgba(0,0,0,0.1);
+  }
+"""
+
+# Gradio interface
+def gradio_interface(user_message):
+    response, sources = cassandre_bot.predict(user_message)
+    return response, sources
+
+# Create Gradio app
+demo = gr.Blocks(css=css)
+
+with demo:
+    gr.HTML("""<h1 style="text-align:center">Cassandre</h1>""")
+    with gr.Row():
+        with gr.Column(scale=2):
+            text_input = gr.Textbox(label="Votre question ou votre instruction", lines=3)
+            text_button = gr.Button("Interroger Cassandre")
+        with gr.Column(scale=3):
+            text_output = gr.HTML(label="La réponse de Cassandre")
+    with gr.Row():
+        embedding_output = gr.HTML(label="Les sources utilisées")
+    
+    text_button.click(gradio_interface, inputs=text_input, outputs=[text_output, embedding_output])
+
+# Launch the app
+if __name__ == "__main__":
+    demo.launch()

gitattributes(7)

@@ -0,0 +1,7 @@
+education_corrected/chroma.sqlite3 filter=lfs diff=lfs merge=lfs -text
+education_corrected/e150eb41-e894-45c4-b97c-80ced9ff2123/data_level0.bin filter=lfs diff=lfs merge=lfs -text
+education_corrected/a9ac8f33-9498-450a-ae99-f116efb66330/data_level0.bin filter=lfs diff=lfs merge=lfs -text
+education_corrected/6af97eb5-0cfa-40b2-a4df-732ca13bd66a/data_level0.bin filter=lfs diff=lfs merge=lfs -text
+content/lancedb_data/eduv1.lance/_indices/fts/55ac048af92d47c0903552a94300d4e3.store filter=lfs diff=lfs merge=lfs -text
+content18/lancedb_data/edunat18.lance/_indices/fts/1655747b9eec413b8f22e287af9a4e8e.store filter=lfs diff=lfs merge=lfs -text
+content19/lancedb_data/edunat19.lance/_indices/fts/28b0ccef3dd3401396a9ab134c426bc6.store filter=lfs diff=lfs merge=lfs -text

gitattributes(8)

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+education_database/chroma.sqlite3 filter=lfs diff=lfs merge=lfs -text

requirements(2).txt

@@ -0,0 +1,12 @@
+transformers
+torch
+einops
+accelerate
+tiktoken
+scipy
+vllm
+lancedb
+sentence_transformers
+gradio
+pandas
+tantivy

theme_builder(2).py

@@ -0,0 +1,3 @@
+import gradio as gr
+
+gr.themes.builder()