- Base set of docs converted to markdown (README, AUTHORS, CONTRIBUTING, HISTORY)
- Introduces the
timeout
keyword argument, which allows the caller to specify a timeout after which requests to the scrapyd server give up. This works as per the underlyingrequests
library, and raisesrequests.exceptions.Timeout
when the timeout is exceeded. See docs for usage.
v2.0.0 shipped with docs which were slightly out of date for the cancel endpoint, this release corrects that.
Why Version 2? This package has been production ready and stable in use for over a year now, so it's ready to commit to a stable API /w semver. Version 1 has deliberately been skipped to make it absolutely clear that this release contains a breaking change:
Breaking changes:
- The cancel job endpoint now returns the previous state of the successfully
cancelled spider rather than a simple boolean True/False. This change was
made because:
a) the boolean return was relatively useless and actually hiding data the
scrapyd API passes us as part of the cancel endpoint response.
b) before this change, the method would have returned
True
only if the cancelled job was previously running, and this resulted in us incorrectly reportingFalse
when a pending job was cancelled. This may require no changes to your codebase but nevertheless it is a change in a public API, thus the requirement for major version bumping.
Other changes:
- The cancel job endpoint now accepts a
signal
keyword argument which is the termination signal Scrapyd uses to cancel the spider job. If not specified, the value is not sent to the scrapyd endpoint at all, therefore allows scrapyd control over which default signal gets used (currentlyTERM
).
- Added the new
job_status
method which can retrieve the job status of a specific job from a project. See docs for usage. - Increased and improved test coverage.
- First release on PyPI.