File size: 19,543 Bytes
ada3f28 | 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 | import logging
import os
from datetime import datetime
from langchain_community.graphs import Neo4jGraph
from src.shared.common_fn import create_gcs_bucket_folder_name_hashed, delete_uploaded_local_file, load_embedding_model
from src.document_sources.gcs_bucket import delete_file_from_gcs
from src.shared.constants import BUCKET_UPLOAD
from src.entities.source_node import sourceNode
import json
class graphDBdataAccess:
def __init__(self, graph: Neo4jGraph):
self.graph = graph
def update_exception_db(self, file_name, exp_msg):
try:
job_status = "Failed"
result = self.get_current_status_document_node(file_name)
is_cancelled_status = result[0]['is_cancelled']
if bool(is_cancelled_status):
job_status = 'Cancelled'
self.graph.query(
"""MERGE(d:Document {fileName :$fName}) SET d.status = $status, d.errorMessage = $error_msg""",
{"fName": file_name, "status": job_status, "error_msg": exp_msg})
except Exception as e:
error_message = str(e)
logging.error(f"Error in updating document node status as failed: {error_message}")
raise Exception(error_message)
def create_source_node(self, obj_source_node: sourceNode):
try:
job_status = "New"
logging.info("creating source node if does not exist")
self.graph.query("""MERGE(d:Document {fileName :$fn}) SET d.fileSize = $fs, d.fileType = $ft ,
d.status = $st, d.url = $url, d.awsAccessKeyId = $awsacc_key_id,
d.fileSource = $f_source, d.createdAt = $c_at, d.updatedAt = $u_at,
d.processingTime = $pt, d.errorMessage = $e_message, d.nodeCount= $n_count,
d.relationshipCount = $r_count, d.model= $model, d.gcsBucket=$gcs_bucket,
d.gcsBucketFolder= $gcs_bucket_folder, d.language= $language,d.gcsProjectId= $gcs_project_id,
d.is_cancelled=False, d.total_chunks=0, d.processed_chunk=0,
d.access_token=$access_token""",
{"fn": obj_source_node.file_name, "fs": obj_source_node.file_size,
"ft": obj_source_node.file_type, "st": job_status,
"url": obj_source_node.url,
"awsacc_key_id": obj_source_node.awsAccessKeyId, "f_source": obj_source_node.file_source,
"c_at": obj_source_node.created_at,
"u_at": obj_source_node.created_at, "pt": 0, "e_message": '', "n_count": 0, "r_count": 0,
"model": obj_source_node.model,
"gcs_bucket": obj_source_node.gcsBucket,
"gcs_bucket_folder": obj_source_node.gcsBucketFolder,
"language": obj_source_node.language, "gcs_project_id": obj_source_node.gcsProjectId,
"access_token": obj_source_node.access_token})
except Exception as e:
error_message = str(e)
logging.info(f"error_message = {error_message}")
self.update_exception_db(self, obj_source_node.file_name, error_message)
raise Exception(error_message)
def update_source_node(self, obj_source_node: sourceNode):
try:
params = {}
if obj_source_node.file_name is not None and obj_source_node.file_name != '':
params['fileName'] = obj_source_node.file_name
if obj_source_node.status is not None and obj_source_node.status != '':
params['status'] = obj_source_node.status
if obj_source_node.created_at is not None:
params['createdAt'] = obj_source_node.created_at
if obj_source_node.updated_at is not None:
params['updatedAt'] = obj_source_node.updated_at
if obj_source_node.processing_time is not None and obj_source_node.processing_time != 0:
params['processingTime'] = round(obj_source_node.processing_time.total_seconds(), 2)
if obj_source_node.node_count is not None:
params['nodeCount'] = obj_source_node.node_count
if obj_source_node.relationship_count is not None:
params['relationshipCount'] = obj_source_node.relationship_count
if obj_source_node.model is not None and obj_source_node.model != '':
params['model'] = obj_source_node.model
if obj_source_node.total_chunks is not None and obj_source_node.total_chunks != 0:
params['total_chunks'] = obj_source_node.total_chunks
if obj_source_node.is_cancelled is not None:
params['is_cancelled'] = obj_source_node.is_cancelled
if obj_source_node.processed_chunk is not None:
params['processed_chunk'] = obj_source_node.processed_chunk
if obj_source_node.retry_condition is not None:
params['retry_condition'] = obj_source_node.retry_condition
param = {"props": params}
print(f'Base Param value 1 : {param}')
query = "MERGE(d:Document {fileName :$props.fileName}) SET d += $props"
logging.info("Update source node properties")
self.graph.query(query, param)
except Exception as e:
error_message = str(e)
self.update_exception_db(self.file_name, error_message)
raise Exception(error_message)
def get_source_list(self):
"""
Args:
uri: URI of the graph to extract
db_name: db_name is database name to connect to graph db
userName: Username to use for graph creation ( if None will use username from config file )
password: Password to use for graph creation ( if None will use password from config file )
file: File object containing the PDF file to be used
model: Type of model to use ('Diffbot'or'OpenAI GPT')
Returns:
Returns a list of sources that are in the database by querying the graph and
sorting the list by the last updated date.
"""
logging.info("Get existing files list from graph")
query = "MATCH(d:Document) WHERE d.fileName IS NOT NULL RETURN d ORDER BY d.updatedAt DESC"
result = self.graph.query(query)
list_of_json_objects = [entry['d'] for entry in result]
return list_of_json_objects
def update_KNN_graph(self):
"""
Update the graph node with SIMILAR relationship where embedding scrore match
"""
index = self.graph.query("""show indexes yield * where type = 'VECTOR' and name = 'vector'""")
# logging.info(f'show index vector: {index}')
knn_min_score = os.environ.get('KNN_MIN_SCORE')
if len(index) > 0:
logging.info('update KNN graph')
self.graph.query("""MATCH (c:Chunk)
WHERE c.embedding IS NOT NULL AND count { (c)-[:SIMILAR]-() } < 5
CALL db.index.vector.queryNodes('vector', 6, c.embedding) yield node, score
WHERE node <> c and score >= $score MERGE (c)-[rel:SIMILAR]-(node) SET rel.score = score
""",
{"score": float(knn_min_score)}
)
else:
logging.info("Vector index does not exist, So KNN graph not update")
def connection_check_and_get_vector_dimensions(self):
"""
Get the vector index dimension from database and application configuration and DB connection status
Args:
uri: URI of the graph to extract
userName: Username to use for graph creation ( if None will use username from config file )
password: Password to use for graph creation ( if None will use password from config file )
db_name: db_name is database name to connect to graph db
Returns:
Returns a status of connection from NEO4j is success or failure
"""
db_vector_dimension = self.graph.query("""SHOW INDEXES YIELD *
WHERE type = 'VECTOR' AND name = 'vector'
RETURN options.indexConfig['vector.dimensions'] AS vector_dimensions
""")
result_chunks = self.graph.query("""match (c:Chunk) return size(c.embedding) as embeddingSize, count(*) as chunks,
count(c.embedding) as hasEmbedding
""")
embedding_model = os.getenv('EMBEDDING_MODEL')
embeddings, application_dimension = load_embedding_model(embedding_model)
logging.info(f'embedding model:{embeddings} and dimesion:{application_dimension}')
# print(chunks_exists)
if self.graph:
if len(db_vector_dimension) > 0:
return {'db_vector_dimension': db_vector_dimension[0]['vector_dimensions'],
'application_dimension': application_dimension, 'message': "Connection Successful"}
else:
if len(db_vector_dimension) == 0 and len(result_chunks) == 0:
logging.info("Chunks and vector index does not exists in database")
return {'db_vector_dimension': 0, 'application_dimension': application_dimension,
'message': "Connection Successful", "chunks_exists": False}
elif len(db_vector_dimension) == 0 and result_chunks[0]['hasEmbedding'] == 0 and result_chunks[0][
'chunks'] > 0:
return {'db_vector_dimension': 0, 'application_dimension': application_dimension,
'message': "Connection Successful", "chunks_exists": True}
else:
return {'message': "Connection Successful"}
def execute_query(self, query, param=None):
return self.graph.query(query, param)
def get_current_status_document_node(self, file_name):
query = """
MATCH(d:Document {fileName : $file_name}) RETURN d.status AS Status , d.processingTime AS processingTime,
d.nodeCount AS nodeCount, d.model as model, d.relationshipCount as relationshipCount,
d.total_chunks AS total_chunks , d.fileSize as fileSize,
d.is_cancelled as is_cancelled, d.processed_chunk as processed_chunk, d.fileSource as fileSource
"""
param = {"file_name": file_name}
return self.execute_query(query, param)
def delete_file_from_graph(self, filenames, source_types, deleteEntities: str, merged_dir: str, uri):
# filename_list = filenames.split(',')
filename_list = list(map(str.strip, json.loads(filenames)))
source_types_list = list(map(str.strip, json.loads(source_types)))
gcs_file_cache = os.environ.get('GCS_FILE_CACHE')
# source_types_list = source_types.split(',')
for (file_name, source_type) in zip(filename_list, source_types_list):
merged_file_path = os.path.join(merged_dir, file_name)
if source_type == 'local file' and gcs_file_cache == 'True':
folder_name = create_gcs_bucket_folder_name_hashed(uri, file_name)
delete_file_from_gcs(BUCKET_UPLOAD, folder_name, file_name)
else:
logging.info(f'Deleted File Path: {merged_file_path} and Deleted File Name : {file_name}')
delete_uploaded_local_file(merged_file_path, file_name)
query_to_delete_document = """
MATCH (d:Document) where d.fileName in $filename_list and d.fileSource in $source_types_list
with collect(d) as documents
unwind documents as d
optional match (d)<-[:PART_OF]-(c:Chunk)
detach delete c, d
return count(*) as deletedChunks
"""
query_to_delete_document_and_entities = """
match (d:Document) where d.fileName IN $filename_list and d.fileSource in $source_types_list
detach delete d
with collect(d) as documents
unwind documents as d
match (d)<-[:PART_OF]-(c:Chunk)
detach delete c
with *
match (c)-[:HAS_ENTITY]->(e)
where not exists { (e)<-[:HAS_ENTITY]-()-[:PART_OF]->(d2) where not d2 in documents }
detach delete e
"""
param = {"filename_list": filename_list, "source_types_list": source_types_list}
if deleteEntities == "true":
result = self.execute_query(query_to_delete_document_and_entities, param)
logging.info(
f"Deleting {len(filename_list)} documents = '{filename_list}' from '{source_types_list}' from database")
else:
result = self.execute_query(query_to_delete_document, param)
logging.info(
f"Deleting {len(filename_list)} documents = '{filename_list}' from '{source_types_list}' with their entities from database")
return result, len(filename_list)
def list_unconnected_nodes(self):
query = """
MATCH (e:!Chunk&!Document)
WHERE NOT exists { (e)--(:!Chunk&!Document) }
OPTIONAL MATCH (doc:Document)<-[:PART_OF]-(c:Chunk)-[:HAS_ENTITY]->(e)
RETURN e {.*, embedding:null, elementId:elementId(e), labels:labels(e)} as e,
collect(distinct doc.fileName) as documents, count(distinct c) as chunkConnections
ORDER BY e.id ASC
LIMIT 100
"""
query_total_nodes = """
MATCH (e:!Chunk&!Document)
WHERE NOT exists { (e)--(:!Chunk&!Document) }
RETURN count(*) as total
"""
nodes_list = self.execute_query(query)
total_nodes = self.execute_query(query_total_nodes)
return nodes_list, total_nodes[0]
def delete_unconnected_nodes(self, unconnected_entities_list):
entities_list = list(map(str.strip, json.loads(unconnected_entities_list)))
query = """
MATCH (e) WHERE elementId(e) IN $elementIds
DETACH DELETE e
"""
param = {"elementIds": entities_list}
return self.execute_query(query, param)
def get_duplicate_nodes_list(self):
score_value = float(os.environ.get('DUPLICATE_SCORE_VALUE'))
text_distance = int(os.environ.get('DUPLICATE_TEXT_DISTANCE'))
query_duplicate_nodes = """
MATCH (n:!Chunk&!Document)
WITH n
WHERE n.embedding IS NOT NULL
AND n.id IS NOT NULL
AND NOT toLower(n.id) CONTAINS 'figure' // Exclude IDs containing 'figure'
WITH n
WITH n ORDER BY count {{ (n)--() }} DESC, size(n.id) DESC // updated
WITH collect(n) as nodes
UNWIND nodes as n
WITH n, [other in nodes
// only one pair, same labels e.g. Person with Person
WHERE elementId(n) < elementId(other) and labels(n) = labels(other)
// at least embedding similarity of X
AND
(
// either contains each other as substrings or has a text edit distinct of less than 3
(size(other.id) > 2 AND toLower(n.id) CONTAINS toLower(other.id)) OR
(size(n.id) > 2 AND toLower(other.id) CONTAINS toLower(n.id))
OR (size(n.id)>5 AND apoc.text.distance(toLower(n.id), toLower(other.id)) < $duplicate_text_distance)
OR
vector.similarity.cosine(other.embedding, n.embedding) > $duplicate_score_value
)] as similar
WHERE size(similar) > 0
// remove duplicate subsets
with collect([n]+similar) as all
CALL {{ with all
unwind all as nodes
with nodes, all
// skip current entry if it's smaller and a subset of any other entry
where none(other in all where other <> nodes and size(other) > size(nodes) and size(apoc.coll.subtract(nodes, other))=0)
return head(nodes) as n, tail(nodes) as similar
}}
OPTIONAL MATCH (doc:Document)<-[:PART_OF]-(c:Chunk)-[:HAS_ENTITY]->(n)
{return_statement}
"""
return_query_duplicate_nodes = """
RETURN n {.*, embedding:null, elementId:elementId(n), labels:labels(n)} as e,
[s in similar | s {.id, .description, labels:labels(s), elementId: elementId(s)}] as similar,
collect(distinct doc.fileName) as documents, count(distinct c) as chunkConnections
ORDER BY e.id ASC
"""
return_query_duplicate_nodes_total = "RETURN COUNT(DISTINCT(n)) as total"
param = {"duplicate_score_value": score_value, "duplicate_text_distance": text_distance}
nodes_list = self.execute_query(query_duplicate_nodes.format(return_statement=return_query_duplicate_nodes),
param=param)
total_nodes = self.execute_query(
query_duplicate_nodes.format(return_statement=return_query_duplicate_nodes_total), param=param)
return nodes_list, total_nodes[0]
def merge_duplicate_nodes(self, duplicate_nodes_list):
nodes_list = json.loads(duplicate_nodes_list)
print(f'Nodes list to merge {nodes_list}')
query = """
UNWIND $rows AS row
CALL { with row
MATCH (first) WHERE elementId(first) = row.firstElementId
MATCH (rest) WHERE elementId(rest) IN row.similarElementIds
WITH first, collect (rest) as rest
WITH [first] + rest as nodes
CALL apoc.refactor.mergeNodes(nodes,
{properties:"discard",mergeRels:true, produceSelfRel:false, preserveExistingSelfRels:false, singleElementAsArray:true})
YIELD node
RETURN size(nodes) as mergedCount
}
RETURN sum(mergedCount) as totalMerged
"""
param = {"rows": nodes_list}
return self.execute_query(query, param)
def drop_create_vector_index(self, isVectorIndexExist):
"""
drop and create the vector index when vector index dimesion are different.
"""
embedding_model = os.getenv('EMBEDDING_MODEL')
embeddings, dimension = load_embedding_model(embedding_model)
if isVectorIndexExist == 'true':
self.graph.query("""drop index vector""")
# self.graph.query("""drop index vector""")
self.graph.query("""CREATE VECTOR INDEX `vector` if not exists for (c:Chunk) on (c.embedding)
OPTIONS {indexConfig: {
`vector.dimensions`: $dimensions,
`vector.similarity_function`: 'cosine'
}}
""",
{
"dimensions": dimension
}
)
return "Drop and Re-Create vector index succesfully"
|