Spaces:
Sleeping
Sleeping
MVPilgrim
commited on
Commit
·
cee962e
1
Parent(s):
18a3a1b
Don't create objs each time.
Browse files- semsearch.py +90 -91
semsearch.py
CHANGED
|
@@ -157,6 +157,7 @@ client = weaviate.WeaviateClient(
|
|
| 157 |
grpc_host="localhost",
|
| 158 |
grpc_port="50051",
|
| 159 |
grpc_secure=False,
|
|
|
|
| 160 |
),
|
| 161 |
# auth_client_secret=weaviate.auth.AuthApiKey("secr3tk3y"),
|
| 162 |
# additional_headers={
|
|
@@ -174,107 +175,105 @@ client.connect()
|
|
| 174 |
#wpCollection = createWebpageCollection()
|
| 175 |
#wpChunkCollection = createChunksCollection()
|
| 176 |
logger.info("#### createWebpageCollection() entered.")
|
| 177 |
-
if client.collections.exists("Documents"):
|
| 178 |
-
client.collections.delete("Documents")
|
| 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 |
-
"dataType": ["text"],
|
| 216 |
-
"description": "HTML page content.",
|
| 217 |
-
"moduleConfig": {
|
| 218 |
-
"text2vec-transformers": {
|
| 219 |
-
"vectorizePropertyName": True,
|
| 220 |
-
"tokenization": "whitespace"
|
| 221 |
}
|
| 222 |
}
|
| 223 |
-
|
| 224 |
-
|
| 225 |
-
|
| 226 |
-
wpCollection = client.collections.create_from_dict(class_obj)
|
| 227 |
|
| 228 |
logger.info("#### createChunksCollection() entered.")
|
| 229 |
-
if client.collections.exists("Chunks"):
|
| 230 |
-
client.collections.delete("Chunks")
|
| 231 |
-
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
|
| 235 |
-
|
| 236 |
-
|
| 237 |
-
|
| 238 |
-
"vectorizeClassName": True
|
| 239 |
-
}
|
| 240 |
-
},
|
| 241 |
-
"vectorIndexType": "hnsw",
|
| 242 |
-
"vectorIndexConfig": {
|
| 243 |
-
"distance": "cosine",
|
| 244 |
-
},
|
| 245 |
-
"properties": [
|
| 246 |
-
{
|
| 247 |
-
"name": "chunk",
|
| 248 |
-
"dataType": ["text"],
|
| 249 |
-
"description": "Single webpage chunk.",
|
| 250 |
-
"vectorizer": "text2vec-transformers",
|
| 251 |
-
"moduleConfig": {
|
| 252 |
-
"text2vec-transformers": {
|
| 253 |
-
"vectorizePropertyName": False,
|
| 254 |
-
"skip": False,
|
| 255 |
-
"tokenization": "lowercase"
|
| 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 |
###########################################################
|
|
|
|
| 157 |
grpc_host="localhost",
|
| 158 |
grpc_port="50051",
|
| 159 |
grpc_secure=False,
|
| 160 |
+
log_level="WARNING"
|
| 161 |
),
|
| 162 |
# auth_client_secret=weaviate.auth.AuthApiKey("secr3tk3y"),
|
| 163 |
# additional_headers={
|
|
|
|
| 175 |
#wpCollection = createWebpageCollection()
|
| 176 |
#wpChunkCollection = createChunksCollection()
|
| 177 |
logger.info("#### createWebpageCollection() entered.")
|
| 178 |
+
if not client.collections.exists("Documents"):
|
| 179 |
+
#client.collections.delete("Documents")
|
| 180 |
+
class_obj = {
|
| 181 |
+
"class": "Documents",
|
| 182 |
+
"description": "For first attempt at loading a Weviate database.",
|
| 183 |
+
"vectorizer": "text2vec-transformers",
|
| 184 |
+
"moduleConfig": {
|
| 185 |
+
"text2vec-transformers": {
|
| 186 |
+
"vectorizeClassName": False
|
| 187 |
+
}
|
| 188 |
+
},
|
| 189 |
+
"vectorIndexType": "hnsw",
|
| 190 |
+
"vectorIndexConfig": {
|
| 191 |
+
"distance": "cosine",
|
| 192 |
+
},
|
| 193 |
+
"properties": [
|
| 194 |
+
{
|
| 195 |
+
"name": "title",
|
| 196 |
+
"dataType": ["text"],
|
| 197 |
+
"description": "HTML doc title.",
|
| 198 |
+
"vectorizer": "text2vec-transformers",
|
| 199 |
+
"moduleConfig": {
|
| 200 |
+
"text2vec-transformers": {
|
| 201 |
+
"vectorizePropertyName": True,
|
| 202 |
+
"skip": False,
|
| 203 |
+
"tokenization": "lowercase"
|
| 204 |
+
}
|
| 205 |
+
},
|
| 206 |
+
"invertedIndexConfig": {
|
| 207 |
+
"bm25": {
|
| 208 |
+
"b": 0.75,
|
| 209 |
+
"k1": 1.2
|
| 210 |
+
},
|
| 211 |
}
|
| 212 |
},
|
| 213 |
+
{
|
| 214 |
+
"name": "content",
|
| 215 |
+
"dataType": ["text"],
|
| 216 |
+
"description": "HTML page content.",
|
| 217 |
+
"moduleConfig": {
|
| 218 |
+
"text2vec-transformers": {
|
| 219 |
+
"vectorizePropertyName": True,
|
| 220 |
+
"tokenization": "whitespace"
|
| 221 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 222 |
}
|
| 223 |
}
|
| 224 |
+
]
|
| 225 |
+
}
|
| 226 |
+
wpCollection = client.collections.create_from_dict(class_obj)
|
|
|
|
| 227 |
|
| 228 |
logger.info("#### createChunksCollection() entered.")
|
| 229 |
+
if not client.collections.exists("Chunks"):
|
| 230 |
+
#client.collections.delete("Chunks")
|
| 231 |
+
class_obj = {
|
| 232 |
+
"class": "Chunks",
|
| 233 |
+
"description": "Collection for document chunks.",
|
| 234 |
+
"vectorizer": "text2vec-transformers",
|
| 235 |
+
"moduleConfig": {
|
| 236 |
+
"text2vec-transformers": {
|
| 237 |
+
"vectorizeClassName": True
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 238 |
}
|
| 239 |
},
|
| 240 |
+
"vectorIndexType": "hnsw",
|
| 241 |
+
"vectorIndexConfig": {
|
| 242 |
+
"distance": "cosine",
|
| 243 |
},
|
| 244 |
+
"properties": [
|
| 245 |
+
{
|
| 246 |
+
"name": "chunk",
|
| 247 |
+
"dataType": ["text"],
|
| 248 |
+
"description": "Single webpage chunk.",
|
| 249 |
+
"vectorizer": "text2vec-transformers",
|
| 250 |
+
"moduleConfig": {
|
| 251 |
+
"text2vec-transformers": {
|
| 252 |
+
"vectorizePropertyName": False,
|
| 253 |
+
"skip": False,
|
| 254 |
+
"tokenization": "lowercase"
|
| 255 |
+
}
|
| 256 |
}
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"name": "chunk_index",
|
| 260 |
+
"dataType": ["int"]
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"name": "webpage",
|
| 264 |
+
"dataType": ["Documents"],
|
| 265 |
+
"description": "Webpage content chunks.",
|
| 266 |
+
|
| 267 |
+
"invertedIndexConfig": {
|
| 268 |
+
"bm25": {
|
| 269 |
+
"b": 0.75,
|
| 270 |
+
"k1": 1.2
|
| 271 |
+
}
|
| 272 |
+
}
|
| 273 |
+
}
|
| 274 |
+
]
|
| 275 |
+
}
|
| 276 |
+
wpChunkCollection = client.collections.create_from_dict(class_obj)
|
| 277 |
|
| 278 |
|
| 279 |
###########################################################
|