Add files using upload-large-folder tool

commands.txt

@@ -0,0 +1,96 @@
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name Fast-F1 --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/Fast-F1 > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_Fast-F1.out
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name Flexget --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/Flexget > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_Flexget.out
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name LLaMA-Factory --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/LLaMA-Factory > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_LLaMA-Factory.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name PyBaMM --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/PyBaMM > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_PyBaMM.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name PyPSA --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/PyPSA > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_PyPSA.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name Radicale --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/Radicale > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_Radicale.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name Solaar --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/Solaar > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_Solaar.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name WeasyPrint --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/WeasyPrint > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_WeasyPrint.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name astroid --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/astroid > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_astroid.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name conda --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/conda > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_conda.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name django --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/django > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_django.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name dspy --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/dspy > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_dspy.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name dvc --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/dvc > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_dvc.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name faker --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/faker > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_faker.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name fastmcp --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/fastmcp > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_fastmcp.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name faststream --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/faststream > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_faststream.out
+
+
+
+
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name feature_engine --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/feature_engine > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_feature_engine.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name filesystem_spec --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/filesystem_spec > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_filesystem_spec.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name flask --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/flask > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_flask.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name fonttools --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/fonttools > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_fonttools.out
+
+
+
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name fusesoc --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/fusesoc > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_fusesoc.out
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name geopandas --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/geopandas > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_geopandas.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name gitingest --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/gitingest > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_gitingest.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name icloud_photos_downloader --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/icloud_photos_downloader > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_icloud_photos_downloader.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name instructlab --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/instructlab > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_instructlab.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name jax --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/jax > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_jax.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name kedro --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/kedro > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_kedro.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name kirara-ai --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/kirara-ai > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_kirara-ai.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name linkding --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/linkding > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_linkding.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name litellm --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/litellm > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_litellm.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name llama-stack --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/llama-stack > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_llama-stack.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name llama_deploy --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/llama_deploy > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_llama_deploy.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name loguru --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/loguru > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_loguru.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name marshmallow --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/marshmallow > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_marshmallow.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name mcp-atlassian --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/mcp-atlassian > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_mcp-atlassian.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name mesa --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/mesa > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_mesa.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name mypy --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/mypy > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_mypy.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name networkx --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/networkx > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_networkx.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name ntc-templates --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/ntc-templates > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_ntc-templates.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name openai-agents-python --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/openai-agents-python > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_openai-agents-python.out
+
+
+
+
+
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name patroni --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/patroni > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_patroni.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name pdm --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/pdm > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_pdm.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name pipenv --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/pipenv > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_pipenv.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name poetry --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/poetry > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_poetry.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name privacyidea --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/privacyidea > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_privacyidea.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name pydicom --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/pydicom > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_pydicom.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name pymdown-extensions --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/pymdown-extensions > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_pymdown-extensions.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name pyomo --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/pyomo > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_pyomo.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name python-control --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/python-control > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_python-control.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name python-telegram-bot --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/python-telegram-bot > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_python-telegram-bot.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name python --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/python > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_python.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name pyvista --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/pyvista > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_pyvista.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name qtile --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/qtile > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_qtile.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name reflex --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/reflex > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_reflex.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name scipy --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/scipy > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_scipy.out
+
+
+
+
+CUDA_VISIBLE_DEVICES=0 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name scrapy-splash --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/scrapy-splash > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_scrapy-splash.out
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name segmentation_models.pytorch --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/segmentation_models.pytorch > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_segmentation_models.pytorch.out
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name sh --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/sh > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_sh.out
+CUDA_VISIBLE_DEVICES=1 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name smart_open --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/smart_open > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_smart_open.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name smolagents --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/smolagents > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_smolagents.out
+CUDA_VISIBLE_DEVICES=2 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name sqlfluff --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/sqlfluff > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_sqlfluff.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name sqllineage --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/sqllineage > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_sqllineage.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name starlette --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/starlette > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_starlette.out
+CUDA_VISIBLE_DEVICES=3 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name streamlink --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/streamlink > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_streamlink.out
+CUDA_VISIBLE_DEVICES=4 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name tablib --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/tablib > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_tablib.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name torchtune --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/torchtune > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_torchtune.out
+CUDA_VISIBLE_DEVICES=5 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name tox --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/tox > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_tox.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name transitions --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/transitions > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_transitions.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name trimesh --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/trimesh > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_trimesh.out
+CUDA_VISIBLE_DEVICES=6 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name twine --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/twine > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_twine.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name urllib3 --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/urllib3 > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_urllib3.out
+CUDA_VISIBLE_DEVICES=7 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name wemake-python-styleguide --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/wemake-python-styleguide > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_wemake-python-styleguide.out
+CUDA_VISIBLE_DEVICES=0 nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name yt-dlp --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/yt-dlp > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_yt-dlp.out
+
+
+
+
+
+
+

find_what_repo_live_need_to_embed.py

@@ -0,0 +1,66 @@
+import os
+import random
+
+def find_missing_repos():
+    repos_dir = "/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/repos"
+    embedding_dir = "/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding"
+    cnt = 0
+    # 遍历repos目录下的所有文件夹
+    for folder_name in os.listdir(repos_dir):
+        folder_path = os.path.join(repos_dir, folder_name)
+        
+        # 确保是文件夹
+        if os.path.isdir(folder_path):
+            A = folder_name
+            
+            # 检查embedding_dir下是否存在A或output_A文件夹
+            path_A = os.path.join(embedding_dir, A)
+            path_output_A = os.path.join(embedding_dir, f"output_{A}")
+            
+            # 如果两者都不存在，则打印输出
+            if not os.path.exists(path_A) and not os.path.exists(path_output_A):
+                # 检查pyggraph目录下是否存在A.timed.pt文件
+                pyggraph_file = f"/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/pyggraph/{A}.timed.pt"
+                if os.path.exists(pyggraph_file):
+                    cnt += 1
+                    print(A)
+                    
+    print(f"Total missing repos: {cnt}")
+    
+def generate_commands():
+    
+    repos_dir = "/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/repos"
+    embedding_dir = "/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding"
+    commands = []
+    
+    # 遍历repos目录下的所有文件夹
+    for folder_name in os.listdir(repos_dir):
+        folder_path = os.path.join(repos_dir, folder_name)
+        
+        # 确保是文件夹
+        if os.path.isdir(folder_path):
+            A = folder_name
+            
+            # 检查embedding_dir下是否存在A或output_A文件夹
+            path_A = os.path.join(embedding_dir, A)
+            path_output_A = os.path.join(embedding_dir, f"output_{A}")
+            
+            # 如果两者都不存在，则生成命令
+            if not os.path.exists(path_A) and not os.path.exists(path_output_A):
+                # 检查pyggraph目录下是否存在A.timed.pt文件
+                pyggraph_file = f"/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/pyggraph/{A}.timed.pt"
+                if os.path.exists(pyggraph_file):
+                    gpu_id = random.randint(0, 7)
+                    command = f"CUDA_VISIBLE_DEVICES={gpu_id} nohup python /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/run_full_generation.py --repo_name {A} --batch_size 8 --output_dir /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/{A} > /data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/logs/node_content_embedding_{A}.out"
+                    commands.append(command)
+    
+    # 写入txt文件
+    with open("/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding/commands.txt", "w") as f:
+        for command in commands:
+            f.write(command + "\n")
+    
+    print(f"Generated {len(commands)} commands and saved to commands.txt")
+
+if __name__ == "__main__":
+    find_missing_repos()
+    generate_commands()

generate_content_embeddings.py

@@ -0,0 +1,382 @@
+
+
+import torch
+import datasets
+import json
+import numpy as np
+from pathlib import Path
+from typing import Dict, List, Tuple
+import argparse
+from tqdm import tqdm
+import gc
+
+try:
+    from safetensors.torch import save_file
+    SAFETENSORS_AVAILABLE = True
+    print("✅ safetensors可用")
+except ImportError:
+    print("❌ safetensors不可用，将使用numpy格式")
+    SAFETENSORS_AVAILABLE = False
+
+# vLLM相关导入
+try:
+    from vllm import LLM
+    from transformers import AutoTokenizer
+    VLLM_AVAILABLE = True
+    print("✅ vLLM可用")
+except ImportError:
+    print("❌ vLLM不可用")
+    VLLM_AVAILABLE = False
+
+class ContentEmbeddingGenerator:
+    
+    
+    def __init__(self, 
+                 repo_name: str = "astropy",
+                 model_path: str = "/data/wangjuntong/FROM_120/data1/.cache/modelscope/hub/models/Qwen/Qwen3-Embedding-8B",
+                 output_dir: str = "./output"):
+        """
+        
+        
+        Args:
+            repo_name: 仓库名称
+            model_path: embedding模型路径
+            output_dir: 输出目录
+        """
+        self.repo_name = repo_name
+        self.model_path = model_path
+        self.output_dir = Path(output_dir)
+        self.output_dir.mkdir(parents=True, exist_ok=True)
+        
+        
+        self.graph_file = f"/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/pyggraph/{repo_name}.timed.pt"
+        self.dataset_dir = f"/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/savedata/repos/{repo_name}/"
+        
+        
+        self.model = None
+        self.tokenizer = None
+        self._load_model()
+        
+    def _load_model(self):
+        """加载vLLM embedding模型"""
+        if not VLLM_AVAILABLE:
+            print("❌ vLLM不可用，无法生成embedding")
+            return
+            
+        try:
+            print(f"🔄 加载embedding模型: {self.model_path}")
+            self.model = LLM(
+                model=self.model_path,
+                task="embed",
+                enforce_eager=True,
+                gpu_memory_utilization=0.9,
+                max_model_len=32768
+            )
+            
+            self.tokenizer = AutoTokenizer.from_pretrained(
+                self.model_path,
+                padding_side='left',
+                trust_remote_code=True
+            )
+            print("✅ 模型加载成功")
+            
+        except Exception as e:
+            print(f"❌ 模型加载失败: {e}")
+            self.model = None
+            self.tokenizer = None
+    
+    def load_graph_data(self) -> torch.Tensor:
+        """加载图数据，获取节点ID列表"""
+        print(f"🔄 加载图数据: {self.graph_file}")
+        
+        try:
+            data = torch.load(self.graph_file, map_location='cpu', weights_only=False)
+            num_nodes = data.num_nodes
+            print(f"✅ 图数据加载成功，节点数: {num_nodes}")
+            
+            
+            node_ids = torch.arange(num_nodes, dtype=torch.long)
+            return node_ids
+            
+        except Exception as e:
+            print(f"❌ 图数据加载失败: {e}")
+            raise
+    
+    def load_dataset(self) -> datasets.Dataset:
+        
+        print(f"🔄 加载原始数据集: {self.dataset_dir}")
+        
+        try:
+            dataset = datasets.Dataset.load_from_disk(self.dataset_dir)
+            print(f"✅ 数据集加载成功，样本数: {len(dataset)}")
+            return dataset
+            
+        except Exception as e:
+            print(f"❌ 数据集加载失败: {e}")
+            raise
+    
+    def extract_node_contents(self, node_ids: torch.Tensor, dataset: datasets.Dataset) -> Dict[int, Dict]:
+        
+        print("🔄 提取节点内容...")
+        
+        node_contents = {}
+        dataset_len = len(dataset)
+        
+        for node_id in tqdm(node_ids, desc="提取节点内容"):
+            node_id_int = int(node_id)
+            
+            if node_id_int < dataset_len:
+                
+                sample = dataset[node_id_int]
+                
+                node_contents[node_id_int] = {
+                    'node_id': node_id_int,
+                    'name': sample.get('name', ''),
+                    'path': sample.get('path', ''),
+                    'attr': sample.get('attr', ''),  
+                    'type': sample.get('type', ''),
+                    'start_commit': sample.get('start_commit', ''),
+                    'end_commit': sample.get('end_commit', '')
+                }
+            else:
+                
+                node_contents[node_id_int] = {
+                    'node_id': node_id_int,
+                    'name': '',
+                    'path': '',
+                    'attr': '',
+                    'type': '',
+                    'start_commit': '',
+                    'end_commit': ''
+                }
+        
+        print(f"✅ 节点内容提取完成，有效节点: {sum(1 for v in node_contents.values() if v['attr'])}")
+        return node_contents
+    
+    def generate_embeddings(self, node_contents: Dict[int, Dict], batch_size: int = 32) -> Dict[int, np.ndarray]:
+        """生成embeddings - 为所有节点生成embedding，包括空节点"""
+        if not self.model:
+            print("❌ 模型未加载，跳过embedding生成")
+            return {}
+        
+        print("🔄 生成embeddings...")
+        
+        
+        texts = []
+        node_id_order = []
+        
+        for node_id in sorted(node_contents.keys()):
+            content = node_contents[node_id]
+            text = content['name'] + content['path'] + content['attr']
+
+            if not text or text.strip() == '{}' or text.strip() == '':
+                text = " "  
+            
+            texts.append(text)
+            node_id_order.append(node_id)
+        
+        print(f"需要生成embedding的文本数量: {len(texts)}")
+        
+        
+        embeddings = {}
+        total_batches = (len(texts) + batch_size - 1) // batch_size
+        
+        for i in tqdm(range(0, len(texts), batch_size), desc="生成embeddings", total=total_batches):
+            batch_texts = texts[i:i + batch_size]
+            batch_node_ids = node_id_order[i:i + batch_size]
+            
+            try:
+                
+                tokenized = self.tokenizer(
+                    batch_texts,
+                    padding=True,
+                    truncation=True,
+                    max_length=32768,
+                    return_tensors="pt"
+                )
+                
+                
+                processed_texts = self.tokenizer.batch_decode(
+                    tokenized["input_ids"], 
+                    skip_special_tokens=True
+                )
+                
+                
+                emb_outputs = self.model.embed(processed_texts)
+                
+                
+                for j, node_id in enumerate(batch_node_ids):
+                    emb_output = emb_outputs[j]
+                    
+                    
+                    if hasattr(emb_output, "embedding"):
+                        embedding = emb_output.embedding
+                    elif hasattr(emb_output, "hidden_states"):
+                        embedding = emb_output.hidden_states
+                    elif hasattr(emb_output, "outputs") and hasattr(emb_output.outputs, "embedding"):
+                        embedding = emb_output.outputs.embedding
+                    elif hasattr(emb_output, "outputs") and hasattr(emb_output.outputs, "hidden_states"):
+                        embedding = emb_output.outputs.hidden_states
+                    else:
+                        raise ValueError("无法从模型输出中提取embedding")
+                    
+                    
+                    if hasattr(embedding, "cpu"):
+                        embedding = embedding.cpu().numpy()
+                    elif hasattr(embedding, "numpy"):
+                        embedding = embedding.numpy()
+                    else:
+                        embedding = np.array(embedding)
+                    
+                    embeddings[node_id] = embedding
+                
+                
+                if torch.cuda.is_available():
+                    torch.cuda.empty_cache()
+                gc.collect()
+                
+            except Exception as e:
+                print(f"❌ 批次 {i//batch_size + 1} embedding生成失败: {e}")
+                
+                for node_id in batch_node_ids:
+                    embeddings[node_id] = np.zeros(4096, dtype=np.float32)
+        
+        print(f"✅ Embedding生成完成，成功生成: {len(embeddings)}")
+        return embeddings
+    
+    def save_content_json(self, node_contents: Dict[int, Dict]):
+        """保存节点内容为JSON"""
+        output_file = self.output_dir / f"{self.repo_name}_node_contents.json"
+        print(f"🔄 保存节点内容到: {output_file}")
+        
+        
+        sorted_contents = {}
+        for node_id in sorted(node_contents.keys()):
+            sorted_contents[str(node_id)] = node_contents[node_id]
+        
+        try:
+            with open(output_file, 'w', encoding='utf-8') as f:
+                json.dump(sorted_contents, f, indent=2, ensure_ascii=False)
+            print(f"✅ 节点内容保存完成")
+            
+        except Exception as e:
+            print(f"❌ 节点内容保存失败: {e}")
+            raise
+    
+    def save_embeddings_safetensor(self, embeddings: Dict[int, np.ndarray], all_node_ids: List[int] = None):
+        """保存embedding为safetensor格式 - 确保按node_id顺序保存"""
+        if not embeddings:
+            print("⚠️ 没有embedding数据，跳过保存")
+            return
+        
+        print("🔄 保存embeddings...")
+        
+        
+        if all_node_ids:
+            for node_id in all_node_ids:
+                if node_id not in embeddings:
+                    print(f"⚠️ 节点 {node_id} 缺少embedding，使用零向量")
+                    embeddings[node_id] = np.zeros(4096, dtype=np.float32)
+        
+        
+        sorted_node_ids = sorted(embeddings.keys())
+        embedding_list = []
+        
+        for node_id in sorted_node_ids:
+            embedding = embeddings[node_id]
+            if isinstance(embedding, np.ndarray):
+                embedding_list.append(torch.from_numpy(embedding))
+            else:
+                embedding_list.append(torch.tensor(embedding))
+        
+        
+        embeddings_tensor = torch.stack(embedding_list, dim=0)
+        print(f"Embeddings tensor shape: {embeddings_tensor.shape}")
+        
+        
+        print(f"节点ID范围: {min(sorted_node_ids)} - {max(sorted_node_ids)}")
+        print(f"节点ID是否连续: {sorted_node_ids == list(range(min(sorted_node_ids), max(sorted_node_ids) + 1))}")
+        
+        
+        if SAFETENSORS_AVAILABLE:
+            output_file = self.output_dir / f"{self.repo_name}_embeddings.safetensors"
+            tensors = {"embeddings": embeddings_tensor}
+            save_file(tensors, output_file)
+            print(f"✅ Safetensor保存至: {output_file}")
+        else:
+            output_file = self.output_dir / f"{self.repo_name}_embeddings.npz"
+            np.savez_compressed(output_file, embeddings=embeddings_tensor.numpy())
+            print(f"✅ Numpy数组保存至: {output_file}")
+        
+        # 保存索引映射
+        index_file = self.output_dir / f"{self.repo_name}_embedding_index.json"
+        index_mapping = {
+            'node_ids': sorted_node_ids,
+            'embedding_dim': embeddings_tensor.shape[1],
+            'num_nodes': len(sorted_node_ids),
+            'format': 'safetensors' if SAFETENSORS_AVAILABLE else 'numpy',
+            'min_node_id': min(sorted_node_ids),
+            'max_node_id': max(sorted_node_ids),
+            'is_continuous': sorted_node_ids == list(range(min(sorted_node_ids), max(sorted_node_ids) + 1))
+        }
+        
+        with open(index_file, 'w', encoding='utf-8') as f:
+            json.dump(index_mapping, f, indent=2)
+        print(f"✅ 索引映射保存至: {index_file}")
+    
+    def run(self, batch_size: int = 32):
+        """运行完整流程"""
+        print(f"🚀 开始处理仓库: {self.repo_name}")
+        
+        try:
+            # 1. 加载图数据获取节点ID
+            node_ids = self.load_graph_data()
+            
+            # 2. 加载原始数据集
+            dataset = self.load_dataset()
+            
+            # 3. 提取节点内容
+            node_contents = self.extract_node_contents(node_ids, dataset)
+            
+            # 4. 保存节点内容JSON
+            #self.save_content_json(node_contents)
+            
+            # 5. 生成embeddings（包括所有节点）
+            embeddings = self.generate_embeddings(node_contents, batch_size)
+            
+            # 6. 保存embeddings，传入所有node_ids确保完整性
+            self.save_embeddings_safetensor(embeddings, all_node_ids=[int(nid) for nid in node_ids])
+            
+            print(f"🎉 处理完成！")
+            print(f"📊 统计信息:")
+            print(f"  - 总节点数: {len(node_contents)}")
+            print(f"  - 有内容节点数: {sum(1 for v in node_contents.values() if v['attr'] and v['attr'] != '{}')}")
+            print(f"  - 生成embedding数: {len(embeddings)}")
+            
+        except Exception as e:
+            print(f"❌ 处理失败: {e}")
+            import traceback
+            traceback.print_exc()
+            raise
+
+def main():
+    parser = argparse.ArgumentParser(description="生成节点内容和embedding")
+    parser.add_argument('--repo_name', type=str, default='astropy', help='仓库名称')
+    parser.add_argument('--model_path', type=str, 
+                       default='/data/wangjuntong/FROM_120/data1/.cache/modelscope/hub/models/Qwen/Qwen3-Embedding-8B',
+                       help='Embedding模型路径')
+    parser.add_argument('--output_dir', type=str, default='./output', help='输出目录')
+    parser.add_argument('--batch_size', type=int, default=1, help='批处理大小')
+    
+    args = parser.parse_args()
+    
+    generator = ContentEmbeddingGenerator(
+        repo_name=args.repo_name,
+        model_path=args.model_path,
+        output_dir=args.output_dir
+    )
+    
+    generator.run(batch_size=args.batch_size)
+
+if __name__ == "__main__":
+    main()

generate_rewriter_embedding_vllm.py

@@ -0,0 +1,125 @@
+"""
+Generate embedding for Queries from Rewriter's Inferer using vLLM and Qwen3-Embedding-8B
+"""
+
+from vllm import LLM
+from transformers import AutoTokenizer
+import torch
+import os
+import tqdm
+import json
+import pickle
+from datasets import Dataset
+
+# Input and output paths
+rewriter_output_path = "/data1/wangjuntong/CodeFuse-CGM-wxy/rewriter_output_post.json"
+rewriter_embedding_path = "rewriter_embedding.pkl"
+
+# Initialize vLLM model and tokenizer
+print("Loading Qwen3-Embedding-8B model...")
+model = LLM(
+    model="Qwen/Qwen3-Embedding-8B",
+    task="embed",
+    trust_remote_code=True
+)
+
+tokenizer = AutoTokenizer.from_pretrained(
+    'Qwen/Qwen3-Embedding-8B',
+    padding_side='left',
+    trust_remote_code=True,
+    local_files_only=os.path.exists('/home/wangjuntong/.cache/modelscope/hub/models/Qwen/Qwen3-Embedding-8B')
+)
+
+max_length = 32768
+
+def get_embeddings(batch):
+    """
+    Generate embeddings for a batch of texts using vLLM
+    """
+    texts = batch["query"]
+    
+    # Tokenize and decode (to ensure proper format)
+    prompt_token_ids = tokenizer(
+        texts,
+        padding=True,
+        truncation=True,
+        max_length=max_length,
+        return_tensors="pt",
+    )
+    
+    prompts = tokenizer.batch_decode(prompt_token_ids["input_ids"], skip_special_tokens=True)
+    
+    # Get embeddings using vLLM
+    emb_outputs = model.embed(prompts)
+    
+    # Extract embeddings
+    embeddings = []
+    for out in emb_outputs:
+        # Extract embedding based on model output structure
+        if hasattr(out, "embedding"):
+            emb = out.embedding
+        elif hasattr(out, "outputs") and hasattr(out.outputs, "embedding"):
+            emb = out.outputs.embedding
+        else:
+            emb = out.outputs[0].embedding if out.outputs else None
+            
+        if emb is None:
+            raise ValueError("Cannot extract embedding from model output")
+            
+        # Convert to list if it's a tensor
+        if hasattr(emb, "tolist"):
+            emb = emb.tolist()
+        elif hasattr(emb, "cpu"):
+            emb = emb.cpu().numpy().tolist()
+        
+        embeddings.append(emb)
+    
+    return {"embedding": embeddings}
+
+if __name__ == "__main__":
+    # Load rewriter output
+    with open(rewriter_output_path, 'r') as file:
+        rewriter_output_dict = json.load(file)
+    
+    # Prepare dataset
+    data = []
+    if isinstance(rewriter_output_dict, dict):
+        for instance_id, item in rewriter_output_dict.items():
+            query = item.get("rewriter_inferer", "")
+            if query:  # Skip empty queries
+                data.append({
+                    "instance_id": instance_id,
+                    "query": query
+                })
+    elif isinstance(rewriter_output_dict, list):
+        for idx, item in enumerate(rewriter_output_dict):
+            query = item.get("rewriter_inferer", "")
+            if query:  # Skip empty queries
+                data.append({
+                    "instance_id": str(idx),
+                    "query": query
+                })
+    
+    # Create dataset
+    dataset = Dataset.from_list(data)
+    
+    # Process in batches using map
+    embedded_dataset = dataset.map(
+        get_embeddings,
+        batched=True,
+        batch_size=10000,  # Adjust batch size based on your GPU memory
+        remove_columns=["query"]
+    )
+    
+    # Create final embedding dictionary
+    query_embedding_dict = {}
+    for item in tqdm.tqdm(embedded_dataset, desc="Organizing embeddings"):
+        instance_id = item["instance_id"]
+        embedding = item["embedding"]
+        query_embedding_dict[instance_id] = embedding
+    
+    # Save embeddings
+    with open(rewriter_embedding_path, 'wb') as f:
+        pickle.dump(query_embedding_dict, f)
+    
+    print(f"Saved query embeddings to {rewriter_embedding_path}")

run_full_generation.py

@@ -0,0 +1,76 @@
+#!/usr/bin/env python3
+
+import sys
+import os
+import argparse
+sys.path.append('/data/wangjuntong/FROM_120/data1/RepoGNN_backup_0809/get_content_embedding')
+
+from generate_content_embeddings import ContentEmbeddingGenerator
+
+def main():
+    parser = argparse.ArgumentParser(description="生成完整的节点内容和embedding")
+    parser.add_argument('--repo_name', type=str, default='astropy', help='仓库名称')
+    parser.add_argument('--batch_size', type=int, default=1, help='批处理大小')
+    parser.add_argument('--output_dir', type=str, default='./output', help='输出目录')
+    parser.add_argument('--max_nodes', type=int, default=None, help='最大节点数（用于测试）')
+    
+    args = parser.parse_args()
+    
+    print(f"🚀 开始生成 {args.repo_name} 的完整节点内容和embedding")
+    print(f"配置参数:")
+    print(f"  - 仓库名称: {args.repo_name}")
+    print(f"  - 批处理大小: {args.batch_size}")
+    print(f"  - 输出目录: {args.output_dir}")
+    print(f"  - 最大节点数: {args.max_nodes or '全部'}")
+    
+    
+    generator = ContentEmbeddingGenerator(
+        repo_name=args.repo_name,
+        output_dir=args.output_dir
+    )
+    
+    try:
+        
+        node_ids = generator.load_graph_data()
+        
+        
+        if args.max_nodes and args.max_nodes < len(node_ids):
+            node_ids = node_ids[:args.max_nodes]
+            print(f"⚠️ 限制节点数为: {len(node_ids)}")
+        
+        
+        dataset = generator.load_dataset()
+        
+        # 3. 提取节点内容
+        node_contents = generator.extract_node_contents(node_ids, dataset)
+        
+        # 统计有效节点
+        valid_nodes = {k: v for k, v in node_contents.items() 
+                      if v['attr'] and v['attr'] != '{}'}
+        print(f"📊 节点统计:")
+        print(f"  - 总节点数: {len(node_contents)}")
+        print(f"  - 有效节点数（有代码内容）: {len(valid_nodes)}")
+        print(f"  - 空节点数: {len(node_contents) - len(valid_nodes)}")
+        
+        # 4. 保存节点内容JSON
+        #generator.save_content_json(node_contents)
+        
+        # 5. 对所有节点生成embedding（包括空节点）
+        print(f"🔄 开始生成 {len(node_contents)} 个节点的embedding（包括空节点）...")
+        embeddings = generator.generate_embeddings(node_contents, batch_size=args.batch_size)
+        
+        # 6. 保存embedding
+        if embeddings:
+            generator.save_embeddings_safetensor(embeddings)
+        else:
+            print("⚠️ 没有找到有效的节点内容，跳过embedding生成")
+        
+        print(f"🎉 处理完成！")
+        
+    except Exception as e:
+        print(f"❌ 处理失败: {e}")
+        import traceback
+        traceback.print_exc()
+
+if __name__ == "__main__":
+    main()