Datasets:

changliu8541
/

assemblage-deephistory

Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 80, in _split_generators
                  raise ValueError(
                  ...<2 lines>...
                  )
              ValueError: The TAR archives of the dataset should be in WebDataset format, but the files in the archive don't share the same prefix or the same types.
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ~~~~~~~~~~~~~~~~~~~~~~~^
                      path=dataset,
                      ^^^^^^^^^^^^^
                      config_name=config,
                      ^^^^^^^^^^^^^^^^^^^
                      token=hf_token,
                      ^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                      path,
                  ...<6 lines>...
                      **config_kwargs,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Assemblage-DeepHistory

A C/C++ binary dataset spanning GCC, Clang and MSVC, multiple optimization levels, Linux and Windows, and multi-year version histories.

Scope

Full database: 279,285 binary records; 194,701 distinct SHA-256 hashes.
Paper library subset: 73,610 records from 248 packages.
CVE cohort: 329 CVEs across 55 packages; 10,196 CVE/binary pairs covering 2,789 binary records, including libraries and executables.

Changelog

2026-09-14 — CVE labeling correction. Corrected labels for 70 CVEs in the database.

2026-05. Migrated metadata to DuckDB and corrected missing or mismatched function-ID associations.

File	Contents
`deephistory.duckdb.tar.zst`	Build, source, function and debug metadata, plus CVE labels. 99.5 GiB compressed; 573 GiB extracted.
`binaries.tar.zst`	Binary payloads under `binaries/`, indexed by `binaries.path`. 201 GiB compressed.

Core tables: binaries, functions, pdbs, rvas, and lines.

CVE table/view	Use
`cve_metadata`	Descriptions, source targets, patches and fix commits.
`cve_binary_labels`	One row per CVE/binary pair: claim, mapping status, scope and eligibility.
`cve_function_mappings`	Individual function locations, native identities and evidence.

cve_label_evidence and cve_label_generations retain detailed provenance.

Label interpretation

has_vulnerability records a version claim: true = affected, false = excluded, null = unknown. Unknowns are not negative examples. Function mappings do not prove exploitability or that a fix is present; all current binary_security_verified values are false.

Inspect mapping_status, localization_scope, and eligibility flags before using labels. Native identities may have function_id=NULL; consult cve_function_mappings. Some provenance references external source, debug or analysis-cache files.

Quick start

Requires zstd. Metadata queries do not require downloading the binary archive.

pip install duckdb huggingface_hub
hf download changliu8541/assemblage-deephistory deephistory.duckdb.tar.zst \
  --repo-type dataset --local-dir .
tar -I zstd -xf deephistory.duckdb.tar.zst

import duckdb

con = duckdb.connect("deephistory.duckdb", read_only=True)
con.execute("""
    SELECT b.path, b.platform, b.version,
           l.security_status, l.mapping_status, l.localization_scope
    FROM cve_binary_labels l
    JOIN binaries b ON b.id = l.binary_id
    WHERE l.cve_id = ?
""", ["CVE-2023-30774"]).fetchall()

License

CC0-1.0. Underlying projects retain their original licenses; see binaries.license for recorded per-binary license information.

Downloads last month: 78

Total file size:

323 GB