Inputs

Every playbook has a section of input loaders – components that read raw data (files, APIs, databases, etc.) and convert it into Python objects for downstream steps.
An input can accept user‑supplied values passed by the params section.

Common input types

• table – extracts data from tabular files (TSV/CSV)
• avro – loads an Avro file (see docs/docs/inputs/avro.md)
• json, csv, sql, etc.

Example – table

The table loader is a good starting point because it demonstrates the typical parameter set required by most inputs. See the full specification in docs/docs/inputs/table.md:

params:
  gafFile:
    type: File
    value: ../../source/go/goa_human.gaf.gz

inputs:
  gafLoad:
    tableLoad:
      path: "{{params.gafFile}}"
      columns:
        - db
        - id
        - symbol
        - qualifier
        - goID
        - reference
        - evidenceCode
        - from
        - aspect
        - name
        - synonym
        - objectType
        - taxon
        - date
        - assignedBy
        - extension
        - geneProduct

When you run the playbook you can override any of these parameters, e.g.:

sifter run gatplaybook.yaml --param gafFile=/tmp/mydata.tsv