Upload folder using huggingface_hub

config.py

@@ -16,6 +16,7 @@ CLAUDE_API_KEY: str = os.environ.get("CLAUDE_API_KEY", "")
 OPENAI_API_KEY: str = os.environ.get("OPENAI_API_KEY", "")
 GROQ_API_KEY: str   = os.environ.get("GROQ_API_KEY", "")
 NVIDIA_API_KEY: str = os.environ.get("NVIDIA_API_KEY", "")
+HF_API_KEY: str     = os.environ.get("HF_API_KEY", "")
 
 # ── Data settings ─────────────────────────────────────────────────────────────
 DEFAULT_YEARS: int = 5

pipeline/peer_selector.py

@@ -12,6 +12,8 @@ from typing import List, Dict, Optional
 import requests
 
 sys.path.insert(0, str(Path(__file__).parent.parent))
+sys.path.insert(0, str(Path(__file__).parent.parent))
+import config
 from config import FMP_API_KEY, FMP_BASE, REQUEST_TIMEOUT
 from pipeline.resolver import CompanyMeta
 from pipeline.fetcher import fetch_peer_snapshot, _fmp_get
@@ -65,6 +67,69 @@ def _fmp_peers(ticker: str) -> List[str]:
     return []
 
 
+def _hf_intelligent_peers(meta: CompanyMeta, log_fn=None) -> List[str]:
+    """Use Hugging Face Serverless API to intelligently select granular peers."""
+    def log(msg):
+        if log_fn: log_fn(msg)
+        
+    hf_key = getattr(config, "HF_API_KEY", "")
+    if not hf_key:
+        return []
+        
+    log("  → Hugging Face API key detected. Using AI for granular sector breakdown and peer selection...")
+    
+    # Try a fast, high-quality instruction model
+    url = "https://api-inference.huggingface.co/models/meta-llama/Meta-Llama-3-8B-Instruct"
+    headers = {
+        "Authorization": f"Bearer {hf_key}",
+        "Content-Type": "application/json"
+    }
+    
+    desc = meta.description if meta.description else "No description available."
+    prompt = (
+        f"You are a top-tier financial analyst. Given the following company, break down its exact sub-sector "
+        f"and identify its 5 closest publicly traded peers.\n"
+        f"Company Name: {meta.name}\n"
+        f"Industry: {meta.industry}\n"
+        f"Description: {desc}\n\n"
+        f"Return ONLY a valid JSON array of their Yahoo Finance ticker symbols (e.g., [\"MSFT\", \"GOOGL\"]). "
+        f"Do not include any other text, markdown, or explanation. Only the JSON array."
+    )
+    
+    payload = {
+        "inputs": prompt,
+        "parameters": {
+            "max_new_tokens": 100,
+            "return_full_text": False,
+            "temperature": 0.1
+        }
+    }
+    
+    try:
+        r = requests.post(url, headers=headers, json=payload, timeout=25)
+        if r.status_code == 200:
+            result = r.json()
+            if isinstance(result, list) and len(result) > 0:
+                text = result[0].get("generated_text", "")
+                
+                # Extract JSON array using regex
+                import re, json
+                match = re.search(r'\[(.*?)\]', text, re.DOTALL)
+                if match:
+                    arr_str = match.group(0)
+                    tickers = json.loads(arr_str)
+                    if isinstance(tickers, list) and all(isinstance(t, str) for t in tickers):
+                        log(f"  ✓ HF AI successfully identified highly specific peers: {tickers}")
+                        return tickers
+            log("  ⚠ HF AI returned invalid format, falling back...")
+        else:
+            log(f"  ⚠ HF API Error {r.status_code}: {r.text}. Falling back...")
+    except Exception as e:
+        log(f"  ⚠ HF AI Request Failed: {e}. Falling back...")
+        
+    return []
+
+
 def select_peers(
     meta: CompanyMeta,
     max_peers: int = 4,
@@ -83,9 +148,11 @@ def select_peers(
 
     log(f"Selecting peers for {meta.name} ({meta.industry}) …")
 
-    # 1. Try FMP peers endpoint
-    peer_tickers = []
-    if FMP_API_KEY and meta.is_listed and ticker != "UNLISTED":
+    # 1. Try Hugging Face Intelligent Peers
+    peer_tickers = _hf_intelligent_peers(meta, log_fn)
+
+    # 2. Try FMP peers endpoint if HF fails
+    if not peer_tickers and FMP_API_KEY and meta.is_listed and ticker != "UNLISTED":
         peer_tickers = _fmp_peers(ticker)
         if peer_tickers:
             log(f"  → FMP suggested {len(peer_tickers)} peers")

web_app/app.py

@@ -170,6 +170,18 @@ gemini_key = st.sidebar.text_input(
     key="inp_gemini"
 )
 
+st.sidebar.markdown("""
+**Hugging Face** *(Serverless API)*  
+<a href='https://huggingface.co/settings/tokens' target='_blank' style='font-size:0.75rem;color:#34D399;'>Get free key → huggingface.co</a>
+""", unsafe_allow_html=True)
+hf_key = st.sidebar.text_input(
+    "Hugging Face API Key", type="password",
+    placeholder="hf_…",
+    value=os.environ.get("HF_API_KEY", ""),
+    help="Used for intelligent peer selection via Hugging Face Models.",
+    key="inp_hf"
+)
+
 # ── PAID OPTIONS ───────────────────────────────────────────────────────
 st.sidebar.markdown("""
 <div style='background:rgba(99,102,241,0.08);border:1px solid rgba(99,102,241,0.25);
@@ -335,6 +347,13 @@ if start_btn:
         config.OPENAI_API_KEY = openai_key
         config.GROQ_API_KEY   = groq_key
         config.NVIDIA_API_KEY = nvidia_key
+        config.HF_API_KEY     = hf_key
+        
+        os.environ["GEMINI_API_KEY"] = gemini_key
+        os.environ["CLAUDE_API_KEY"] = claude_key
+        os.environ["OPENAI_API_KEY"] = openai_key
+        os.environ["NVIDIA_API_KEY"] = nvidia_key
+        os.environ["HF_API_KEY"]     = hf_key
         
         pipeline.analyst.GEMINI_API_KEY = gemini_key
         pipeline.analyst.CLAUDE_API_KEY = claude_key