Skip to main content

Fuzzy matching utilities for scholarly metadata

Project description

fuzzycat (wip)

Fuzzy matching publications for fatcat.

Example Run

Run any clustering algorithm.

$ time python -m fuzzycat cluster -t tsandcrawler < data/sample10m.json | \
    zstd -c9 > sample_cluster.json.zst
2020-11-18 00:19:48.194 DEBUG __main__ - run_cluster:
    {"key_fail": 0, "key_ok": 9999938, "key_empty": 62, "key_denylist": 0, "num_clusters": 9040789}

real    75m23.045s
user    95m14.455s
sys     3m39.121s

Run verification.

$ time zstdcat -T0 sample_cluster.json.zst | python -m fuzzycat verify > sample_verify.txt

real    7m56.713s
user    8m50.703s
sys     0m29.262s

Example results over 10M docs:

{
  "miss.appendix": 176,
  "miss.blacklisted": 12124,
  "miss.blacklisted_fragment": 9,
  "miss.book_chapter": 46733,
  "miss.component": 2173,
  "miss.contrib_intersection_empty": 73592,
  "miss.dataset_doi": 30806,
  "miss.num_diff": 1,
  "miss.release_type": 19767,
  "miss.short_title": 16737,
  "miss.subtitle": 11975,
  "miss.title_filename": 87,
  "miss.year": 123288,
  "ok.arxiv_version": 90726,
  "ok.dummy": 106196,
  "ok.preprint_published": 10495,
  "ok.slug_title_author_match": 47285,
  "ok.title_author_match": 65685,
  "ok.tokenized_authors": 7592,
  "skip.container_name_blacklist": 20,
  "skip.publisher_blacklist": 456,
  "skip.too_large": 7430,
  "skip.unique": 8808462,
  "total": 9481815
}

A full run

Single threaded, 42h.

$ time zstdcat -T0 release_export_expanded.json.zst | \
    TMPDIR=/bigger/tmp python -m fuzzycat cluster --tmpdir /bigger/tmp -t tsandcrawler | \
    zstd -c9 > cluster_tsandcrawler.json.zst
{
  "key_fail": 0,
  "key_ok": 154202433,
  "key_empty": 942,
  "key_denylist": 0,
  "num_clusters": 124321361
}

real    2559m7.880s
user    2605m41.347s
sys     118m38.141s

So, 29881072 (about 20%) docs in the potentially duplicated set.

Verification (about 15h):

$ time zstdcat -T0 cluster_tsandcrawler.json.zst | python -m fuzzycat verify | \
    zstd -c9 > cluster_tsandcrawler_verified_3c7378.tsv.zst

...

real    927m28.631s
user    939m32.761s
sys     36m47.602s

Use cases

  • take a release entity database dump as JSON lines and cluster releases (according to various algorithms)
  • take cluster information and run a verification step (misc algorithms)
  • create a dataset that contains grouping of releases under works
  • command line tools to generate cache keys, e.g. to match reference strings to release titles (this needs some transparent setup, e.g. filling of a cache before ops)

Usage

Release clusters start with release entities json lines.

$ cat data/sample.json | python -m fuzzycat cluster -t title > out.json

Clustering 1M records (single core) takes about 64s (15K docs/s).

$ head -1 out.json
{
  "k": "裏表紙",
  "v": [
    ...
  ]
}

Using GNU parallel to make it faster.

$ cat data/sample.json | parallel -j 8 --pipe --roundrobin python -m fuzzycat.main cluster -t title

Interestingly, the parallel variants detects fewer clusters (because data is split and clusters are searched within each batch). TODO(miku): sort out sharding bug.

QA

10M release dataset

Notes on cadd28a version clustering (nysiis) and verification.

  • 10M docs
  • 9040789 groups
  • 665447 verification pairs
3578378 OK.TITLE_AUTHOR_MATCH
2989618 Miss.CONTRIB_INTERSECTION_EMPTY
2731528 OK.SLUG_TITLE_AUTHOR_MATCH
2654787 Miss.YEAR
2434532 OK.WORK_ID
2050468 OK.DUMMY
1619330 Miss.SHARED_DOI_PREFIX
1145571 Miss.BOOK_CHAPTER
1023925 Miss.DATASET_DOI
 934075 OK.DATACITE_RELATED_ID
 868951 OK.DATACITE_VERSION
 704154 OK.FIGSHARE_VERSION
 682784 Miss.RELEASE_TYPE
 607117 OK.TOKENIZED_AUTHORS
 298928 OK.PREPRINT_PUBLISHED
 270658 Miss.SUBTITLE
 227537 Miss.SHORT_TITLE
 196402 Miss.COMPONENT
 163158 Miss.CUSTOM_PREFIX_10_5860_CHOICE_REVIEW
 122614 Miss.CUSTOM_PREFIX_10_7916
  79687 OK.CUSTOM_IEEE_ARXIV
  69648 OK.PMID_DOI_PAIR
  46649 Miss.CUSTOM_PREFIX_10_14288
  38598 OK.CUSTOM_BSI_UNDATED
  15465 OK.DOI
  13393 Miss.CUSTOM_IOP_MA_PATTERN
  10378 Miss.CONTAINER
   3045 Miss.BLACKLISTED
   2504 Miss.BLACKLISTED_FRAGMENT
   1574 Miss.TITLE_FILENAME
   1273 Miss.APPENDIX
    104 Miss.NUM_DIFF
      4 OK.ARXIV_VERSION

Cases

Authors may be messy:

A DOI prefix (10.1210, The Endocrine Society) may choose to include the same document in different publications:

Sometimes, a lexicon entry is a "dataset", sometimes a "book", e.g.:

Possible fixes

100 examples

  • accuracy at around 0.8
  • while the results look ok, the reasons are not always the ones that stand out the most (while checking manually)
78 [x]
11 [o]
11 [ ]

Ok cases are now in verify.csv.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

fuzzycat-0.1.7.tar.gz (72.6 kB view details)

Uploaded Source

Built Distribution

fuzzycat-0.1.7-py3-none-any.whl (70.6 kB view details)

Uploaded Python 3

File details

Details for the file fuzzycat-0.1.7.tar.gz.

File metadata

  • Download URL: fuzzycat-0.1.7.tar.gz
  • Upload date:
  • Size: 72.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.2.0 pkginfo/1.6.1 requests/2.25.0 setuptools/50.2.0 requests-toolbelt/0.9.1 tqdm/4.54.1 CPython/3.7.8

File hashes

Hashes for fuzzycat-0.1.7.tar.gz
Algorithm Hash digest
SHA256 158bc80a63489252d19cd22ea58af4d41917784d47f97836238d0859d415ec7d
MD5 611b69f4625a92b38c069bdb57870ea1
BLAKE2b-256 c60ecdc2dfee0b304f629288b40112d48374ca3d0db4973507cf2ce90df7cf41

See more details on using hashes here.

File details

Details for the file fuzzycat-0.1.7-py3-none-any.whl.

File metadata

  • Download URL: fuzzycat-0.1.7-py3-none-any.whl
  • Upload date:
  • Size: 70.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.2.0 pkginfo/1.6.1 requests/2.25.0 setuptools/50.2.0 requests-toolbelt/0.9.1 tqdm/4.54.1 CPython/3.7.8

File hashes

Hashes for fuzzycat-0.1.7-py3-none-any.whl
Algorithm Hash digest
SHA256 5cbb5560c1fdbee83f7ae78b3f3d7dbfc2e8756afb6fbee082aa14a9a737baea
MD5 daa3d7f9c43797a92a64adb8e948a14c
BLAKE2b-256 048022be08d63aaa9abd0b709f299d917acf94b4a3971ac6e66be3f42f3a2699

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page