Spaces:
Sleeping
Sleeping
File size: 19,915 Bytes
133a630 f9ade9b 133a630 f9ade9b 60c3ccb 133a630 28e5413 60c3ccb 133a630 f9ade9b 133a630 f9ade9b 133a630 f9ade9b 133a630 f9ade9b 133a630 a3cafd5 133a630 a3cafd5 133a630 a3cafd5 133a630 f9ade9b 133a630 60c3ccb 133a630 60c3ccb f9ade9b 60c3ccb 133a630 60c3ccb f9ade9b 133a630 60c3ccb 7f47e42 60c3ccb 7f47e42 cabd6cc 133a630 7f47e42 60c3ccb 7f47e42 60c3ccb 133a630 7f47e42 60c3ccb 7f47e42 60c3ccb 7f47e42 133a630 60c3ccb 7f47e42 60c3ccb 7f47e42 60c3ccb 133a630 f9ade9b 133a630 a3cafd5 28e5413 f9ade9b 28e5413 a3cafd5 f9ade9b a3cafd5 133a630 f9ade9b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 | import os
import sys
import logging
import traceback
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
os.environ.setdefault("OPENBLAS_NUM_THREADS", "1")
os.environ.setdefault("OMP_NUM_THREADS", "1")
import json as json_module
import numpy as np
import pandas as pd
from flask import Flask, render_template, jsonify, request, Response, stream_with_context
from utils.helpers import load_data, get_user_preferences
from recommender.collaborative import CollaborativeFiltering
from recommender.content_based import ContentBasedRecommender
from recommender.knowledge_based import KnowledgeBasedRecommender
from recommender.evaluation import Evaluator
from recommender.explainer import Explainer
logging.basicConfig(
level=logging.INFO,
format="%(asctime)s [%(levelname)s] %(message)s",
datefmt="%Y-%m-%d %H:%M:%S",
)
logger = logging.getLogger(__name__)
app = Flask(__name__)
app.secret_key = os.environ.get(
"FLASK_SECRET_KEY",
os.urandom(24).hex()
)
products, users, ratings = load_data()
cf = CollaborativeFiltering(ratings)
cb = ContentBasedRecommender(products)
kb = KnowledgeBasedRecommender(products)
explainer = Explainer(products, users)
evaluator = Evaluator(ratings)
TRAIN = ratings.sample(frac=0.8, random_state=42)
TEST = ratings.drop(TRAIN.index)
cf_train = CollaborativeFiltering(TRAIN)
USER_IDS = sorted(users["user_id"].tolist())
USER_OPTIONS = []
for _, u in users.iterrows():
USER_OPTIONS.append({
"id": int(u["user_id"]),
"name": u["name"],
"age": int(u["age"]),
"categories": u["preferred_categories"].split(",") if isinstance(u["preferred_categories"], str) else [],
"budget_min": float(u["budget_min"]),
"budget_max": float(u["budget_max"]),
"brands": u["favorite_brands"].split(",") if isinstance(u["favorite_brands"], str) else [],
})
CATEGORIES = sorted(products["category"].unique().tolist())
BRANDS = sorted(products["brand"].unique().tolist())
APPROACHES = {
"cf": {
"label": "Collaborative Filtering",
"icon": "π€",
"methods": [
{"id": "user_based", "label": "User-Based CF"},
{"id": "item_based", "label": "Item-Based CF"},
{"id": "svd", "label": "SVD (Matrix Factorization)"},
{"id": "knn", "label": "KNN-Based CF"},
{"id": "slope_one", "label": "Slope One"},
],
},
"content": {
"label": "Content-Based",
"icon": "π·οΈ",
"methods": [
{"id": "tfidf", "label": "TF-IDF Similarity"},
{"id": "feature_match", "label": "Feature Matching"},
],
},
"knowledge": {
"label": "Knowledge-Based",
"icon": "βοΈ",
"methods": [
{"id": "constraint", "label": "Constraint-Based"},
{"id": "rule", "label": "Rule-Based"},
{"id": "utility", "label": "Utility-Based"},
],
},
}
if os.environ.get("PREWARM_CACHE", "0") == "1":
import threading
logger.info("Pre-warming SVD and Slope One caches...")
threading.Thread(
target=lambda: (
list(cf_train.train_svd_generator()),
list(cf_train.compute_slope_one_dev_generator()),
),
daemon=True,
).start()
def get_product_info(product_id):
row = products[products["product_id"] == product_id]
if row.empty:
return None
row = row.iloc[0]
return {
"id": int(row["product_id"]),
"name": row["name"],
"category": row["category"],
"subcategory": row["subcategory"],
"brand": row["brand"],
"price": float(row["price"]),
"avg_rating": float(row["avg_rating"]),
"num_reviews": int(row["num_reviews"]),
}
@app.route("/health")
def health():
return jsonify({"status": "ok", "users": len(USER_OPTIONS), "products": len(products)})
@app.route("/")
def index():
return render_template("index.html",
active_page="home",
users=USER_OPTIONS,
categories=CATEGORIES,
brands=BRANDS,
approaches=APPROACHES)
@app.route("/recommend")
def recommend_page():
return render_template("recommend.html",
active_page="recommend",
users=USER_OPTIONS,
categories=CATEGORIES,
brands=BRANDS,
approaches=APPROACHES)
@app.route("/evaluate")
def evaluate_page():
return render_template("evaluation.html",
active_page="evaluate",
users=USER_OPTIONS,
categories=CATEGORIES,
brands=BRANDS,
approaches=APPROACHES)
@app.route("/api/users")
def api_users():
return jsonify(USER_OPTIONS)
@app.route("/api/user/<int:user_id>")
def api_user(user_id):
prefs = get_user_preferences(users, user_id)
return jsonify(prefs)
@app.route("/api/products")
def api_products():
cat = request.args.get("category")
if cat:
filtered = products[products["category"] == cat]
else:
filtered = products
results = []
for _, row in filtered.iterrows():
results.append(get_product_info(row["product_id"]))
return jsonify(results)
def get_user_rated_items(user_id):
user_ratings = ratings[ratings["user_id"] == user_id]
return user_ratings[user_ratings["rating"] >= 3.5]["product_id"].tolist()
@app.route("/api/recommend", methods=["POST"])
def api_recommend():
data = request.json
user_id = data.get("user_id")
approach = data.get("approach")
method = data.get("method")
n_recs = data.get("n", 10)
if not user_id or not approach or not method:
return jsonify({"error": "Missing required parameters"}), 400
user_rated = get_user_rated_items(user_id)
prefs = get_user_preferences(users, user_id)
try:
if approach == "cf":
recs = cf.recommend(method, user_id, n_recommendations=n_recs)
explanations = []
for pid, score in recs:
details = {"sim_score": score, "count": 10}
explanation = explainer.explain_cf(method, user_id, pid, details)
product = get_product_info(pid)
explanations.append({**product, "score": round(score, 4), "explanation": explanation})
elif approach == "content":
recs = cb.recommend(method, user_profile_items=user_rated, preferences=prefs, n_recommendations=n_recs)
explanations = []
for pid, score in recs:
details = {"score": score}
explanation = explainer.explain_content(method, user_id, pid, details)
product = get_product_info(pid)
explanations.append({**product, "score": round(score, 4), "explanation": explanation})
elif approach == "knowledge":
constraints = {
"budget_min": prefs.get("budget_min", 0),
"budget_max": prefs.get("budget_max", 999999),
"category": list(prefs.get("preferred_categories", set())),
"brand": list(prefs.get("favorite_brands", set())),
}
context = {
"interacted_category": "",
"preferred_categories": prefs.get("preferred_categories", set()),
"budget_min": prefs.get("budget_min", 0),
"budget_max": prefs.get("budget_max", 999999),
"favorite_brands": prefs.get("favorite_brands", set()),
}
recs = kb.recommend(method, constraints=constraints, context=context,
preferences=prefs, n_recommendations=n_recs)
explanations = []
for pid, score in recs:
details = {"score": score, "budget_max": prefs.get("budget_max", 0), "trigger_item": ""}
explanation = explainer.explain_knowledge(method, user_id, pid, details)
product = get_product_info(pid)
explanations.append({**product, "score": round(score, 4), "explanation": explanation})
else:
return jsonify({"error": f"Unknown approach: {approach}"}), 400
return jsonify({"recommendations": explanations})
except Exception as e:
logger.error("Recommendation error: %s\n%s", e, traceback.format_exc())
return jsonify({"error": "An internal error occurred while generating recommendations"}), 500
CF_METHOD_NAMES = ["user_based", "item_based", "svd", "knn", "slope_one"]
CF_METHOD_LABELS = {
"user_based": "User-Based",
"item_based": "Item-Based",
"svd": "SVD",
"knn": "KNN",
"slope_one": "Slope One",
}
@app.route("/api/evaluate/cf/<method>")
def api_evaluate_cf(method):
if method not in CF_METHOD_NAMES:
return jsonify({"error": f"Unknown CF method: {method}"}), 400
try:
result = evaluator.evaluate_cf_method(method, cf_train, TEST, k=5)
return jsonify(result)
except Exception as e:
logger.error("CF evaluation error [%s]: %s\n%s", method, e, traceback.format_exc())
return jsonify({"method": method, "error": "Evaluation failed"})
@app.route("/api/evaluate/cf/<method>/stream")
def api_evaluate_cf_stream(method):
if method not in ("svd", "slope_one"):
return jsonify({"error": f"Streaming not supported for {method}"}), 400
def generate():
try:
if method == "svd":
gen = cf_train.train_svd_generator()
if gen is not None:
for epoch, total in gen:
yield json_module.dumps({"type": "progress", "current": epoch, "total": total}) + "\n"
elif method == "slope_one":
gen = cf_train.compute_slope_one_dev_generator()
if gen is not None:
for item, total in gen:
yield json_module.dumps({"type": "progress", "current": item, "total": total}) + "\n"
yield json_module.dumps({"type": "phase", "label": "Evaluating users..."}) + "\n"
result = evaluator.evaluate_cf_method(method, cf_train, TEST, k=5)
yield json_module.dumps({"type": "result", "data": result}) + "\n"
except Exception as e:
logger.error("Stream evaluation error [%s]: %s\n%s", method, e, traceback.format_exc())
yield json_module.dumps({"type": "error", "message": "Evaluation failed"}) + "\n"
return Response(stream_with_context(generate()), mimetype="application/x-ndjson")
@app.route("/api/evaluate/approaches")
def api_evaluate_approaches():
try:
evaluator.set_test_ratings(TEST)
test_users = TEST["user_id"].unique()[:20]
def approach_precision_recall(recommender_fn):
precisions, recalls = [], []
for uid in test_users:
try:
recs = recommender_fn(uid)
except Exception:
recs = []
rec_items = [r[0] for r in recs]
relevant = evaluator._get_relevant_for_user(uid)
if relevant:
precisions.append(evaluator.precision_at_k(rec_items, relevant, 5))
recalls.append(evaluator.recall_at_k(rec_items, relevant, 5))
return precisions, recalls
def cf_recommender(uid):
return cf_train.recommend("item_based", uid, n_recommendations=10)
train_ratings = ratings[~ratings.index.isin(TEST.index)]
def cb_recommender(uid):
profile = train_ratings[
(train_ratings["user_id"] == uid) & (train_ratings["rating"] >= 3.5)
]["product_id"].tolist()
return cb.recommend("tfidf", user_profile_items=profile, n_recommendations=10)
def kb_recommender(uid):
prefs = get_user_preferences(users, uid)
constraints = {
"budget_min": prefs.get("budget_min", 0),
"budget_max": prefs.get("budget_max", 999999),
"category": list(prefs.get("preferred_categories", set())),
"brand": list(prefs.get("favorite_brands", set())),
}
return kb.recommend("constraint", constraints=constraints, n_recommendations=10)
results = []
cf_p, cf_r = approach_precision_recall(cf_recommender)
if cf_p:
results.append({
"approach": "Collaborative Filtering",
"Precision@5": round(np.mean(cf_p), 4),
"Recall@5": round(np.mean(cf_r), 4),
})
cb_p, cb_r = approach_precision_recall(cb_recommender)
if cb_p:
results.append({
"approach": "Content-Based",
"Precision@5": round(np.mean(cb_p), 4),
"Recall@5": round(np.mean(cb_r), 4),
})
kb_p, kb_r = approach_precision_recall(kb_recommender)
if kb_p:
results.append({
"approach": "Knowledge-Based",
"Precision@5": round(np.mean(kb_p), 4),
"Recall@5": round(np.mean(kb_r), 4),
})
best = max(results, key=lambda a: a.get("Precision@5", 0))["approach"] if results else None
return jsonify({"approaches": results, "best_approach": best})
except Exception as e:
logger.error("Approach evaluation error: %s\n%s", e, traceback.format_exc())
return jsonify({"error": "Approach comparison failed"}), 500
@app.route("/api/products/filter")
def api_products_filter():
cat = request.args.get("category")
brand = request.args.get("brand")
price_min = request.args.get("price_min", type=float)
price_max = request.args.get("price_max", type=float)
q = request.args.get("q", "").lower()
filtered = products.copy()
if cat:
filtered = filtered[filtered["category"] == cat]
if brand:
filtered = filtered[filtered["brand"] == brand]
if price_min is not None:
filtered = filtered[filtered["price"] >= price_min]
if price_max is not None:
filtered = filtered[filtered["price"] <= price_max]
if q:
filtered = filtered[filtered["name"].str.lower().str.contains(q, na=False)]
results = []
for _, row in filtered.iterrows():
results.append(get_product_info(row["product_id"]))
return jsonify({
"total": len(results),
"products": results,
})
@app.route("/api/user/<int:user_id>/preferences", methods=["PUT"])
def api_update_preferences(user_id):
data = request.json
user_idx = users[users["user_id"] == user_id].index
if user_idx.empty:
return jsonify({"error": "User not found"}), 404
if "budget_min" in data:
users.at[user_idx[0], "budget_min"] = data["budget_min"]
if "budget_max" in data:
users.at[user_idx[0], "budget_max"] = data["budget_max"]
prefs = get_user_preferences(users, user_id)
for u in USER_OPTIONS:
if u["id"] == user_id:
u["budget_min"] = float(prefs.get("budget_min", 0))
u["budget_max"] = float(prefs.get("budget_max", 999999))
break
return jsonify({"success": True, "preferences": {
"budget_min": prefs.get("budget_min", 0),
"budget_max": prefs.get("budget_max", 0),
"name": prefs.get("name", ""),
"age": prefs.get("age", 0),
}})
@app.route("/api/users", methods=["POST"])
def api_create_user():
global users
data = request.json
new_id = int(users["user_id"].max() + 1)
new_name = data.get("name", f"User_{new_id}")
new_row = pd.DataFrame([{
"user_id": new_id,
"name": new_name,
"age": int(data.get("age", 25)),
"preferred_categories": ",".join(data.get("categories", [])),
"favorite_brands": ",".join(data.get("brands", [])),
"budget_min": float(data.get("budget_min", 0)),
"budget_max": float(data.get("budget_max", 500)),
}])
users = pd.concat([users, new_row], ignore_index=True)
USER_OPTIONS.append({
"id": new_id,
"name": new_name,
"age": int(data.get("age", 25)),
"categories": data.get("categories", []),
"brands": data.get("brands", []),
"budget_min": float(data.get("budget_min", 0)),
"budget_max": float(data.get("budget_max", 500)),
})
logger.info("Created user %s (ID %d)", new_name, new_id)
return jsonify({"success": True, "user_id": new_id, "name": new_name})
@app.route("/htmx/recommend", methods=["POST"])
def htmx_recommend():
data = request.json or request.form
user_id = data.get("user_id", type=int)
approach = data.get("approach", "cf")
method = data.get("method", "user_based")
n_recs = data.get("n", 10, type=int)
if not user_id:
return '<div class="empty-state"><div class="empty-icon">β οΈ</div><p>Please select a user first.</p></div>'
user_rated = get_user_rated_items(user_id)
prefs = get_user_preferences(users, user_id)
try:
if approach == "cf":
recs = cf.recommend(method, user_id, n_recommendations=n_recs)
elif approach == "content":
recs = cb.recommend(method, user_profile_items=user_rated, preferences=prefs, n_recommendations=n_recs)
elif approach == "knowledge":
constraints = {
"budget_min": prefs.get("budget_min", 0),
"budget_max": prefs.get("budget_max", 999999),
"category": list(prefs.get("preferred_categories", set())),
"brand": list(prefs.get("favorite_brands", set())),
}
recs = kb.recommend(method, constraints=constraints, preferences=prefs, n_recommendations=n_recs)
else:
return f'<div class="empty-state"><div class="empty-icon">β</div><p>Unknown approach: {approach}</p></div>'
except Exception as e:
logger.error("HTMX recommend error: %s\n%s", e, traceback.format_exc())
return f'<div class="empty-state"><div class="empty-icon">β</div><p>An error occurred</p></div>'
if not recs:
return '<div class="empty-state"><div class="empty-icon">π</div><p>No recommendations found.</p></div>'
html = '<div class="product-grid">'
for pid, score in recs:
product = get_product_info(pid)
if not product:
continue
explanation = "Recommended based on your preferences."
html += f'''
<div class="product-card">
<div class="product-icon">{get_category_icon(product["category"])}</div>
<div class="product-name">{product["name"]}</div>
<div class="product-meta">{product["brand"]} Β· {product["subcategory"]}</div>
<div class="compact-row">
<div class="product-price">${product["price"]:.2f}</div>
<div class="product-rating">{stars_html(product["avg_rating"])} {product["avg_rating"]}</div>
</div>
<div class="product-explanation">{explanation}</div>
</div>'''
html += '</div>'
return html
def get_category_icon(category):
icons = {
"Electronics": "π»", "Clothing": "π", "Home & Kitchen": "π ",
"Books": "π", "Sports": "β½", "Beauty": "π", "Toys": "π§Έ", "Automotive": "π"
}
return icons.get(category, "π¦")
def stars_html(rating):
f = int(rating)
return "β
" * f + "β" * (5 - f)
if __name__ == "__main__":
debug_mode = os.environ.get("FLASK_DEBUG", "0") == "1"
port = int(os.environ.get("PORT", 7860))
logger.info("Starting TasteEngine on port %d (debug=%s)", port, debug_mode)
app.run(debug=debug_mode, host="0.0.0.0", port=port)
|