Parallel Data Enrichment

Data Enrichment

Enrich: $ARGUMENTS

Before starting

Inform the user that enrichment may take several minutes depending on the number of rows and fields requested.

Optional: Suggest output columns

If the user gave a vague intent (“enrich these companies with useful info”) and you’re not sure what columns to add, ask the API for a suggestion before kicking off the run:

parallel-cli enrich suggest "Find CEO and recent funding info" --json

The response is an envelope: {title, processor, enriched_columns, warnings}. Extract just the enriched_columns array (not the whole envelope) and pass it as the value of --enriched-columns on enrich run, in place of --intent — the two flags are alternative ways to specify what to enrich, not combined. If suggest returned a processor, pass it through explicitly via --processor on the run call (it’s a tuned recommendation for the schema). Skip this whole section if the user already specified the fields they want.

enrich suggest requires parallel-cli ≥ 0.3.0. If it errors with anything resembling no such command / No such command / unknown command, do not bail — skip the suggestion step, fall through to step 1 with --intent, complete the run, and mention parallel-cli update (or pipx upgrade parallel-web-tools) in the final response so the user picks up the feature next time.

Step 1: Start the enrichment

Use ONE of these command patterns (substitute user’s actual data):

For inline data:

parallel-cli enrich run --data '[{"company": "Google"}, {"company": "Microsoft"}]' --intent "CEO name and founding year" --target "output.csv" --no-wait --json

For CSV file:

parallel-cli enrich run --source-type csv --source "input.csv" --target "output.csv" --source-columns '[{"name": "company", "description": "Company name"}]' --intent "CEO name and founding year" --no-wait --json

If this is a follow-up to a previous research task and you have its interaction_id, add context chaining:

parallel-cli enrich run --data '...' --intent "..." --target "output.csv" --no-wait --json --previous-interaction-id "$INTERACTION_ID"

The enrichment will run with the full context of that prior research — so you can enrich entities discovered earlier without restating what was already found. Note: enrichment does not itself produce a new interaction_id, so you cannot chain a further follow-up off of an enrichment.

IMPORTANT: Always include --no-wait so the command returns immediately instead of blocking.

Parse the --json output to extract taskgroup_id and url. The output is {taskgroup_id, url, num_runs} — there is no interaction_id field, do not look for one. Immediately tell the user:

• Enrichment has been kicked off
• The monitoring URL where they can track progress

Tell them they can background the polling step to continue working while it runs.

Step 2: Poll for results

Pick a concrete output path (e.g., /tmp/enrichment-acme.json). Note: the file is JSON regardless of the extension you choose — it’s an array of {input, output} objects, not a CSV. Name it .json to avoid confusing yourself or the user.

parallel-cli enrich poll "$TASKGROUP_ID" --timeout 540 --output "/tmp/enrichment-<descriptive-name>.json"

Important:

• Use --timeout 540 (9 minutes) to stay within tool execution limits
• The --target from step 1 is unused in --no-wait mode — only --output here determines where results are saved, and the file is always JSON

If the poll times out

Enrichment of large datasets can take longer than 9 minutes. If the poll exits without completing:

1. Tell the user the enrichment is still running server-side
2. Re-run the same parallel-cli enrich poll command to continue waiting

Response format

After step 1: Share the monitoring URL (for tracking progress).

After step 2:

1. Report number of rows enriched
2. Preview first few rows from the output file (it’s a JSON array of {input, output} objects)
3. Tell the user the full path to the output file

Do NOT re-share the monitoring URL after completion — the results are in the output file.

Setup

If parallel-cli is not found, install and authenticate:

/parallel:parallel-cli-setup

If any parallel-cli enrich command returns 403, tell the user balance is likely required. Offer to run parallel-cli balance get, and if needed ask for explicit confirmation before running parallel-cli balance add <amount_cents>. Then retry the original enrichment command.