Spaces:
Sleeping
Sleeping
add fuzzy
Browse files
app.py
CHANGED
|
@@ -6,11 +6,14 @@ from sentence_transformers import SentenceTransformer
|
|
| 6 |
from qdrant_client import QdrantClient
|
| 7 |
from qdrant_client.models import Filter, FieldCondition, MatchValue
|
| 8 |
import os
|
|
|
|
| 9 |
|
| 10 |
|
| 11 |
qdrant_client = QdrantClient(
|
| 12 |
-
url=os.environ.get("Qdrant_url"),
|
| 13 |
-
api_key=os.environ.get("Qdrant_api"),
|
|
|
|
|
|
|
| 14 |
)
|
| 15 |
|
| 16 |
# โมเดลที่โหลดล่วงหน้า
|
|
@@ -39,6 +42,16 @@ model_config = {
|
|
| 39 |
# Global memory to hold feedback state
|
| 40 |
latest_query_result = {"query": "", "result": "", "model": ""}
|
| 41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
# 🌟 Main search function
|
| 43 |
def search_product(query, model_name):
|
| 44 |
start_time = time.time()
|
|
@@ -46,35 +59,39 @@ def search_product(query, model_name):
|
|
| 46 |
if model_name not in model_config:
|
| 47 |
return "❌ ไม่พบโมเดล"
|
| 48 |
|
| 49 |
-
|
|
|
|
|
|
|
|
|
|
| 50 |
collection_name = model_config[model_name]["collection"]
|
| 51 |
-
|
| 52 |
|
| 53 |
-
# Query Qdrant
|
| 54 |
try:
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
|
|
|
| 63 |
except Exception as e:
|
| 64 |
-
|
| 65 |
|
| 66 |
elapsed = time.time() - start_time
|
| 67 |
|
| 68 |
-
|
| 69 |
-
|
|
|
|
|
|
|
|
|
|
| 70 |
result_summary = ""
|
| 71 |
for res in result:
|
| 72 |
line = f"- {res.payload.get('name', '')} (score: {res.score:.4f})"
|
| 73 |
output += line + "\n"
|
| 74 |
result_summary += line + " | "
|
| 75 |
|
| 76 |
-
|
| 77 |
-
latest_query_result["query"] = query
|
| 78 |
latest_query_result["result"] = result_summary.strip()
|
| 79 |
latest_query_result["model"] = model_name
|
| 80 |
|
|
|
|
| 6 |
from qdrant_client import QdrantClient
|
| 7 |
from qdrant_client.models import Filter, FieldCondition, MatchValue
|
| 8 |
import os
|
| 9 |
+
from symspellpy.symspellpy import SymSpell, Verbosity
|
| 10 |
|
| 11 |
|
| 12 |
qdrant_client = QdrantClient(
|
| 13 |
+
#url=os.environ.get("Qdrant_url"),
|
| 14 |
+
#api_key=os.environ.get("Qdrant_api"),
|
| 15 |
+
url=userdata.get("Qdrant_url"),
|
| 16 |
+
api_key=userdata.get("Qdrant_api"),
|
| 17 |
)
|
| 18 |
|
| 19 |
# โมเดลที่โหลดล่วงหน้า
|
|
|
|
| 42 |
# Global memory to hold feedback state
|
| 43 |
latest_query_result = {"query": "", "result": "", "model": ""}
|
| 44 |
|
| 45 |
+
symspell = SymSpell(max_dictionary_edit_distance=2)
|
| 46 |
+
symspell.load_dictionary("symspell_dict_pythainlp.txt", term_index=0, count_index=1)
|
| 47 |
+
|
| 48 |
+
# แก้คำผิด
|
| 49 |
+
def correct_query_with_symspell(query: str) -> str:
|
| 50 |
+
suggestions = symspell.lookup_compound(query, Verbosity.CLOSEST, max_edit_distance=2)
|
| 51 |
+
if suggestions:
|
| 52 |
+
return suggestions[0].term
|
| 53 |
+
return query
|
| 54 |
+
|
| 55 |
# 🌟 Main search function
|
| 56 |
def search_product(query, model_name):
|
| 57 |
start_time = time.time()
|
|
|
|
| 59 |
if model_name not in model_config:
|
| 60 |
return "❌ ไม่พบโมเดล"
|
| 61 |
|
| 62 |
+
# ✨ แทรกขั้นตอน fuzzy correction
|
| 63 |
+
corrected_query = correct_query_with_symspell(query)
|
| 64 |
+
|
| 65 |
+
query_embed = model_config[model_name]["func"](corrected_query)
|
| 66 |
collection_name = model_config[model_name]["collection"]
|
|
|
|
| 67 |
|
|
|
|
| 68 |
try:
|
| 69 |
+
result = qdrant_client.query_points(
|
| 70 |
+
collection_name=collection_name,
|
| 71 |
+
query=query_embed.tolist(),
|
| 72 |
+
with_payload=True,
|
| 73 |
+
query_filter=Filter(
|
| 74 |
+
must=[FieldCondition(key="type", match=MatchValue(value="product"))]
|
| 75 |
+
),
|
| 76 |
+
limit=10
|
| 77 |
+
).points
|
| 78 |
except Exception as e:
|
| 79 |
+
return f"❌ Qdrant error: {str(e)}"
|
| 80 |
|
| 81 |
elapsed = time.time() - start_time
|
| 82 |
|
| 83 |
+
output = f"⏱ Time: {elapsed:.2f}s\n"
|
| 84 |
+
if corrected_query != query:
|
| 85 |
+
output += f"🔧 แก้คำค้นจาก: `{query}` → `{corrected_query}`\n\n"
|
| 86 |
+
output += f"📦 ผลลัพธ์:\n"
|
| 87 |
+
|
| 88 |
result_summary = ""
|
| 89 |
for res in result:
|
| 90 |
line = f"- {res.payload.get('name', '')} (score: {res.score:.4f})"
|
| 91 |
output += line + "\n"
|
| 92 |
result_summary += line + " | "
|
| 93 |
|
| 94 |
+
latest_query_result["query"] = corrected_query
|
|
|
|
| 95 |
latest_query_result["result"] = result_summary.strip()
|
| 96 |
latest_query_result["model"] = model_name
|
| 97 |
|