MCPcopy Create free account
hub / github.com/Alibaba-NLP/ViDoRAG / load_nodes

Method load_nodes

search_engine.py:120–144  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

118 return nodes2dict(recall_results)
119
120 def load_nodes(self):
121 files = os.listdir(self.node_dir)
122 parsed_files = []
123 max_workers = 10
124 if max_workers == 1:
125 for file in tqdm(files):
126 input_file = os.path.join(self.node_dir, file)
127 suffix = input_file.split('.')[-1]
128 if suffix != 'node':
129 continue
130 nodes = nodefile2node(input_file)
131 parsed_files.extend(nodes)
132 else:
133 def parse_file(file,node_dir):
134 input_file = os.path.join(node_dir, file)
135 suffix = input_file.split('.')[-1]
136 if suffix != 'node':
137 return []
138 return nodefile2node(input_file)
139 with ThreadPoolExecutor(max_workers=max_workers) as executor:
140 # results = list(tqdm(executor.map(parse_file, files, self.node_dir), total=len(files)))
141 results = list(tqdm(executor.map(parse_file, files, [self.node_dir]*len(files)), total=len(files)))
142 for result in results:
143 parsed_files.extend(result)
144 return parsed_files
145
146 def load_query_engine(self):
147 print('Loading nodes...')

Callers 1

load_query_engineMethod · 0.95

Calls 1

nodefile2nodeFunction · 0.90

Tested by

no test coverage detected