Add splitting and compression at ungoliant runtime #343
Workflow file for this run
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: get_test_data | |
| on: | |
| push: | |
| branches: [main, dev] | |
| pull_request: | |
| branches: [main, dev] | |
| env: | |
| CARGO_TERM_COLOR: always | |
| jobs: | |
| cache_test_data: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v2 | |
| - name: Cache language identification models | |
| id: cache-lid | |
| uses: actions/cache@v3 | |
| with: | |
| path: lid.*.bin | |
| key: ${{ runner.os }}-lids | |
| - name: Cache CC shards | |
| id: cache-shards | |
| uses: actions/cache@v3 | |
| with: | |
| path: res/shards/*.gz | |
| key: ${{ runner.os }}-shards | |
| - name: Cache blocklist | |
| id: cache-blocklist | |
| uses: actions/cache@v3 | |
| with: | |
| path: res/blocklist | |
| key: ${{ runner.os }}-blocklist | |
| - name: Fetch identification bins | |
| if: steps.cache-lid.outputs.cache-hit != 'true' | |
| run: | | |
| wget https://dl.fbaipublicfiles.com/fasttext/supervised-models/lid.176.bin | |
| wget https://dl.fbaipublicfiles.com/fasttext/supervised-models/lid.189.bin | |
| - name: Fetch blocklist | |
| if: steps.cache-blocklist.outputs.cache-hit != 'true' | |
| run: | | |
| mkdir -p res/blocklist | |
| wget https://github.com/olbat/ut1-blacklists/archive/refs/heads/master.zip | |
| unzip master.zip | |
| mv ut1-blacklists-master/blacklists/* res/blocklist | |
| gzip -d res/blocklist/adult/domains.gz #adult blocklist is compressed | |
| rm -r ut1-blacklists-master | |
| - name: Fetch CC shards | |
| if: steps.cache-shards.outputs.cache-hit != 'true' | |
| run: | | |
| mkdir -p res/shards | |
| wget -O res/shards/0.txt.gz https://data.commoncrawl.org/crawl-data/CC-MAIN-2022-33/segments/1659882570651.49/wet/CC-MAIN-20220807150925-20220807180925-00000.warc.wet.gz | |
| wget -O res/shards/1.txt.gz https://data.commoncrawl.org/crawl-data/CC-MAIN-2022-33/segments/1659882570651.49/wet/CC-MAIN-20220807150925-20220807180925-00001.warc.wet.gz | |
| - name: Install KenLM dependencies | |
| run: sudo apt install -y libboost-all-dev libeigen3-dev | |
| - name: Get sample KenLM model | |
| run: | | |
| mkdir -p res/kenlm | |
| wget -O res/kenlm/en.arpa https://raw.githubusercontent.com/agatan/ctclib/main/data/overfit.arpa | |
| - name: Build | |
| run: cargo build --verbose | |
| - name: Create test directories | |
| run: | | |
| mkdir res/corpus | |
| mkdir res/rebuilt | |
| mkdir -p res/corpus/rebuild | |
| ls res/ | |
| - name: Run tests | |
| run: RUST_BACKTRACE=1 cargo test --verbose | |
| - name: Run cargo-tarpaulin | |
| uses: actions-rs/[email protected] | |
| with: | |
| version: 0.22.0 | |
| timeout: 180 | |
| args: "--avoid-cfg-tarpaulin" | |
| continue-on-error: true | |
| - name: Upload to codecov.io | |
| uses: codecov/codecov-action@v3 |