Chowda

CLAMS processing app

App

Application available at chowda.wgbh-mla.org

Documentation

Visit the documentation for install and usage instructions.

Basic Pipeline with 2 CLAMS Apps

Actors:

• User
• Chowda
  • Webapp
  • Pulbicly available on internet
• DB
  • Postgresql
  • VPC
  • Sperate Cluster?
• Runner
  • Pipeline Runner
  • In CLAMS cluster
• CLAMS
  • Individual CLAMS apps
  • In CLAMS cluster
  • Deployed as either:
    • Webservice
    • Kubernetes Job

Call Sequence Diagram

• User starts job
• Chowda enters it in the DB
• Runner sees new job and starts CLAM1 with initial MMIF
• CLAM1 Processes data and returns MMIF to Runner
• Runner Updates DB and starts CLAM2
• CLAM2 Processes data and returns MMIF to Runner
• Runner Updates DB
• User gets data from Chowda

sequenceDiagram
    actor User
    participant Chowda
    participant DB
    participant Runner
    participant CLAM1
    participant CLAM2


    %% User starts job
    activate User
    note left of User: Hey, Chowda! Start batch 123 on pipeline 456
    User ->> + Chowda: Start batch 123 with files [] through pipeline 456
    Chowda ->> + DB: INSERT (batch, pipeline, Initial mmif) to TABLE pipeline_runs
    DB ->> - Chowda: {pipeline_run_id, status: waiting}
    Chowda ->> - User: Pipeline run {id} is {status}
    deactivate User
    loop backgroud process
        Runner ->> DB: Check for new jobs
    end

    %% Runner
    activate Runner
    note left of Runner: Found new job!
    Runner ->> + CLAM1: Start app 1, with initial MMIF
    deactivate Runner

    %% CLAM1
    note right of CLAM1: Processing!
    loop Status check
        Runner ->> CLAM1: Status report!
        CLAM1 ->> Runner: XX %
        Runner ->> DB: UPDATE pipeline_run_id SET status CLAM1 XX %
    end
    note right of CLAM1: Done!
    CLAM1 ->> - Runner: intermediate MMIF
    activate Runner
    Runner ->> DB: UPDATE pipeline_run_id SET status 'done' + MMIF

    %% CLAM 2
    Runner ->> + CLAM2: Start app 2 with intermediate MMIF
    deactivate Runner
    note right of CLAM2: Processing!
    loop Status check
        Runner ->> CLAM2: Status report!
        CLAM2 ->> Runner: XX %
        Runner ->> DB: UPDATE pipeline_run_id SET status CLAM2 XX %
    end

    %% User update
    activate User
    note left of User: Is it done yet?
    User ->> + Chowda: Pipeline update?
    Chowda ->> + DB: WATCH pipeline_runs WHERE status=running
    DB ->> Chowda: ACTIVE_PIPELINE_RUNS
    Chowda -->> User: Websocket: Batch 123 status [CLAM1: done, CLAM2: XX %]
    note left of User: Not quite, but it should be done soon.

    %% CLAMS finished
    note right of CLAM2: Done!
    CLAM2 -->> - Runner: finished MMIF
    activate Runner
    Runner ->> - DB: UPDATE pipeline_run_id SET status MMIF
    DB ->> Chowda: Batch 123 update
    Chowda -->> User: WS: batch 123 Finished!
    deactivate Chowda
    deactivate DB

    note left of User: Yay! My metadata is perfect!
    deactivate User

develop

pre-commit secret scanning

0. Install ggshield

pip install ggshield
# or
brew install gitguardian/tap/ggshield

1. Login to gitguardian

ggshield auth login

2. Install the pre-commit hooks

pre-commit install