Upload legislation_manager.py

legislation_manager.py

@@ -0,0 +1,153 @@
+import gradio as gr
+import json
+import hashlib
+import os
+import requests
+from bs4 import BeautifulSoup
+from datetime import datetime
+from huggingface_hub import HfApi
+
+DATA_FILE = "/home/ubuntu/legislation_rules.json"
+
+def load_data():
+    if os.path.exists(DATA_FILE):
+        with open(DATA_FILE, 'r') as f:
+            try:
+                return json.load(f)
+            except:
+                return []
+    return []
+
+def save_data(data):
+    with open(DATA_FILE, 'w') as f:
+        json.dump(data, f, indent=2)
+
+def get_canonical_hash(text):
+    return hashlib.sha256(text.strip().encode('utf-8')).hexdigest()
+
+def add_rule_manually(act, title, text, source):
+    data = load_data()
+    det_id = get_canonical_hash(text)
+    
+    if any(r['deterministic_id'] == det_id for r in data):
+        return "Error: This rule already exists in the dataset (matching hash)."
+    
+    new_rule = {
+        "act": act,
+        "section_title": title,
+        "text": text,
+        "source_url": source,
+        "deterministic_id": det_id,
+        "added_at": datetime.now().isoformat()
+    }
+    data.append(new_rule)
+    save_data(data)
+    return f"Successfully added: {title} from {act}"
+
+def scrape_fca_prin():
+    """
+    Scrapes the 12 Principles for Businesses from the FCA Handbook.
+    """
+    url = "https://handbook.fca.org.uk/handbook/PRIN/2/1.html"
+    try:
+        response = requests.get(url)
+        soup = BeautifulSoup(response.content, 'html.parser')
+        
+        principles_table = soup.find('table')
+        if not principles_table:
+            return "Error: Could not find the Principles table on the FCA site."
+        
+        rows = principles_table.find_all('tr')
+        added_count = 0
+        for row in rows:
+            cols = row.find_all('td')
+            if len(cols) >= 2:
+                title = cols[0].get_text(strip=True)
+                text = cols[1].get_text(strip=True)
+                status = add_rule_manually("FCA Handbook: PRIN", f"Principle {title}", text, url)
+                if "Successfully" in status:
+                    added_count += 1
+        
+        return f"Successfully ingested {added_count} FCA Principles."
+    except Exception as e:
+        return f"FCA Scraping failed: {str(e)}"
+
+def sync_to_huggingface(token, dataset_id):
+    if not token or not dataset_id:
+        return "Error: HF Token and Dataset ID are required."
+    
+    api = HfApi()
+    try:
+        api.upload_file(
+            path_or_fileobj=DATA_FILE,
+            path_in_repo="legislation_rules.json",
+            repo_id=dataset_id,
+            repo_type="dataset",
+            token=token
+        )
+        return f"Successfully synced dataset to {dataset_id}"
+    except Exception as e:
+        return f"Sync failed: {str(e)}"
+
+def view_dataset_stats():
+    data = load_data()
+    if not data:
+        return "Dataset is currently empty."
+    
+    stats = f"Total Rules: {len(data)}\n"
+    acts = {}
+    for r in data:
+        acts[r['act']] = acts.get(r['act'], 0) + 1
+    
+    for act, count in acts.items():
+        stats += f"- {act}: {count} rules\n"
+    return stats
+
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# ⚖️ Legislation & FCA Dataset Manager")
+    gr.Markdown("### Create and Expand your Deterministic Rules of Law Dataset")
+    
+    with gr.Tab("➕ Add New Rule"):
+        with gr.Row():
+            with gr.Column():
+                act_input = gr.Textbox(label="Act/Legislation Name", placeholder="e.g., Consumer Rights Act 2015")
+                title_input = gr.Textbox(label="Section/Rule Title", placeholder="e.g., Section 9: Satisfactory Quality")
+                source_input = gr.Textbox(label="Source URL", placeholder="https://www.legislation.gov.uk/...")
+                text_input = gr.TextArea(label="Rule Text", placeholder="Paste the official text here...")
+                add_btn = gr.Button("Add to Deterministic Dataset", variant="primary")
+            with gr.Column():
+                add_status = gr.Textbox(label="Status")
+                stats_view = gr.Textbox(label="Dataset Stats", value=view_dataset_stats())
+                refresh_btn = gr.Button("Refresh Stats")
+
+    with gr.Tab("🏦 FCA Guidelines"):
+        gr.Markdown("### 🛠️ FCA Handbook Automation")
+        gr.Markdown("Click below to automatically ingest the 12 Principles for Businesses (PRIN) from the official FCA Handbook.")
+        fca_btn = gr.Button("Ingest FCA PRIN Principles", variant="secondary")
+        fca_status = gr.Textbox(label="FCA Ingestion Status")
+
+    with gr.Tab("☁️ Sync to Hugging Face"):
+        gr.Markdown("Push your local collection to your Hugging Face account.")
+        hf_token = gr.Textbox(label="HF Write Token", type="password")
+        hf_id = gr.Textbox(label="Dataset ID", placeholder="username/my-legal-dataset")
+        sync_btn = gr.Button("Sync Now")
+        sync_status = gr.Textbox(label="Sync Status")
+
+    add_btn.click(
+        fn=add_rule_manually,
+        inputs=[act_input, title_input, text_input, source_input],
+        outputs=add_status
+    )
+    
+    fca_btn.click(fn=scrape_fca_prin, outputs=fca_status)
+    
+    refresh_btn.click(fn=view_dataset_stats, outputs=stats_view)
+    
+    sync_btn.click(
+        fn=sync_to_huggingface,
+        inputs=[hf_token, hf_id],
+        outputs=sync_status
+    )
+
+if __name__ == "__main__":
+    demo.launch()