Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringlengths
2
17
P10
P1000
P10001
P10003
P10006
P10007
P10008
P1001
P10012
P10013
P10016
P10017
P10018
P1002
P10020
P10021
P10022
P10023
P10027
P10028
P10029
P1003
P10030
P10031
P10032
P10033
P10035
P10037
P10038
P1004
P10040
P10041
P10043
P10045
P10046
P10047
P1005
P10050
P10051
P10052
P10053
P10056
P10058
P10059
P1006
P10060
P10063
P10064
P10065
P10066
P10067
P10068
P10069
P1007
P10072
P10073
P10075
P10076
P10077
P10080
P10082
P10085
P10086
P10090
P10091
P10092
P10096
P10097
P10098
P10099
P101
P1010
P10100
P10106
P10107
P10109
P1011
P10110
P10112
P10113
P10116
P10119
P1012
P10126
P1013
P10131
P10132
P10133
P10134
P10135
P10137
P10138
P10139
P1014
P10140
P10141
P10143
P10144
P10146
P1015
End of preview. Expand in Data Studio

ReflectiVA metadata shard: Mac work package

This package fetches the second half of the referenced Wikidata metadata used by the local ReflectiVA 100K fact-store builder. It does not contain the final fact store. Batch numbering is global and deterministic, so its output can be merged with the Windows-produced first half.

Assigned shard

Referenced IDs: 527,399
Batch size:     50
Total batches:  10,548
Mac range:      [5,274, 10,548)  # 5,274 through 10,547

Use a different public egress IP from the Windows machine. Verify it first:

curl https://api.ipify.org

Download and run on macOS

hf download Kyle-han/reflectiva-metadata-shard \
  --repo-type dataset \
  --local-dir ./reflectiva_metadata_mac

cd ./reflectiva_metadata_mac
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

python fetch_metadata_shard.py \
  --ids referenced_ids.txt \
  --output-dir ./work/cache/metadata \
  --batch-start 5274 \
  --batch-end 10548 \
  --batch-size 50 \
  --workers 2 \
  --archive-output ./reflectiva_metadata_batches_005274_010547.tar.gz \
  2>&1 | tee reflectiva_metadata_mac.log

The command is resumable. Running it again validates and skips completed batch_*.json.gz files. Start with two workers; increase only if 429 responses remain infrequent.

Upload the completed archive

hf auth login
hf upload Kyle-han/reflectiva-metadata-shard \
  ./reflectiva_metadata_batches_005274_010547.tar.gz \
  reflectiva_metadata_batches_005274_010547.tar.gz \
  --repo-type dataset

hf upload Kyle-han/reflectiva-metadata-shard \
  ./reflectiva_metadata_mac.log \
  reflectiva_metadata_mac.log \
  --repo-type dataset

Do not upload the work/cache/metadata directory as thousands of individual files. Upload the single archive generated after all 5,274 batches validate.

Windows merge

After downloading the completed archive on Windows, extract it into the cache parent. The archive contains a top-level metadata/ directory.

tar -xzf reflectiva_metadata_batches_005274_010547.tar.gz `
  -C D:\kb_fact_store_100k\reflectiva\cache

The merged directory should contain every global batch from batch_000000.json.gz through batch_010547.json.gz before final-store building.

Downloads last month
8