Logging and Verbosity

dapi reports what it is doing through Python’s standard logging module, under the dapi logger namespace. By default you see concise milestones; turn the detail up or down globally with one call.

Levels¶

import dapi

dapi.set_log_level("DEBUG")  # full trace of every operation
dapi.set_log_level("INFO")  # default: concise milestones
dapi.set_log_level("WARNING")  # problems only
dapi.set_log_level("QUIET")  # silence dapi entirely

Or set it before import with an environment variable:

export DAPI_LOG_LEVEL=WARNING

What each level shows:

Example — the same read-only-source prepare_inputs call:

# INFO (default)
'/home/jupyter/CommunityData/...' is not writable (e.g. CommunityData); staging to /tmp/dapi-staging-.../DS_input_staged
Bundled 12 files into /tmp/dapi-staging-.../DS_input_staged/tmpSimCenter.zip
Uploading staged files to tapis://designsafe.storage.default/kks32/dapi-staging/DS_input_staged (Tapis files API)
Uploaded 1 file(s); staged_dir is tapis://designsafe.storage.default/kks32/dapi-staging/DS_input_staged

At DEBUG the run also traces each path translation and per-file transfer; at QUIET it prints nothing.

Notebook and terminal behavior¶

The dapi handler writes to stdout, not stderr — so log lines render as normal output in Jupyter (no red boxes) and stay in order with your prints and the job-monitoring progress bars. The same configuration works identically in a terminal.

For applications embedding dapi¶

dapi never touches the root logger. All output flows through the dapi logger hierarchy (dapi.jobs, dapi.files, dapi.simcenter, ...), so standard logging configuration applies:

import logging

logging.getLogger("dapi").setLevel(logging.WARNING)  # quiet dapi only
logging.getLogger("dapi.files").setLevel(logging.DEBUG)  # trace one module

Functions whose purpose is output — job.print_runtime_summary(), ds.jobs.interpret_status(), ds.apps.find(..., verbose=True) — print directly regardless of log level, since you called them to see their report.