Merge branch 'HKUDS:main' into main

This commit is contained in:
jin
2024-11-27 15:16:28 +08:00
committed by GitHub
6 changed files with 313 additions and 23 deletions

View File

@@ -1,6 +1,7 @@
import asyncio
import json
import re
from tqdm.asyncio import tqdm as tqdm_async
from typing import Union
from collections import Counter, defaultdict
import warnings
@@ -342,11 +343,15 @@ async def extract_entities(
)
return dict(maybe_nodes), dict(maybe_edges)
# use_llm_func is wrapped in ascynio.Semaphore, limiting max_async callings
results = await asyncio.gather(
*[_process_single_content(c) for c in ordered_chunks]
)
print() # clear the progress bar
results = []
for result in tqdm_async(
asyncio.as_completed([_process_single_content(c) for c in ordered_chunks]),
total=len(ordered_chunks),
desc="Extracting entities from chunks",
unit="chunk",
):
results.append(await result)
maybe_nodes = defaultdict(list)
maybe_edges = defaultdict(list)
for m_nodes, m_edges in results:
@@ -354,18 +359,38 @@ async def extract_entities(
maybe_nodes[k].extend(v)
for k, v in m_edges.items():
maybe_edges[tuple(sorted(k))].extend(v)
all_entities_data = await asyncio.gather(
*[
_merge_nodes_then_upsert(k, v, knowledge_graph_inst, global_config)
for k, v in maybe_nodes.items()
]
)
all_relationships_data = await asyncio.gather(
*[
_merge_edges_then_upsert(k[0], k[1], v, knowledge_graph_inst, global_config)
for k, v in maybe_edges.items()
]
)
logger.info("Inserting entities into storage...")
all_entities_data = []
for result in tqdm_async(
asyncio.as_completed(
[
_merge_nodes_then_upsert(k, v, knowledge_graph_inst, global_config)
for k, v in maybe_nodes.items()
]
),
total=len(maybe_nodes),
desc="Inserting entities",
unit="entity",
):
all_entities_data.append(await result)
logger.info("Inserting relationships into storage...")
all_relationships_data = []
for result in tqdm_async(
asyncio.as_completed(
[
_merge_edges_then_upsert(
k[0], k[1], v, knowledge_graph_inst, global_config
)
for k, v in maybe_edges.items()
]
),
total=len(maybe_edges),
desc="Inserting relationships",
unit="relationship",
):
all_relationships_data.append(await result)
if not len(all_entities_data):
logger.warning("Didn't extract any entities, maybe your LLM is not working")
return None