1.10.0

New

• New Data Splitter Sink.

  The new DataSplitterSink allows users to extract branches of the data tree and dynamically create files with that data using other parts of the data tree as ID.

  Check the documentation for more information.

• Adds new field total_errors to Valgrind sources.

  Now the Valgrind XML sources will have a new field called total_errors. The idea of this field is to have a count of the amount of errors that occurred. Users won't need to calculate the length of the errors list manually.

• Allow to save the journal in a specific location.

  The journal is a JSON file that reports all details about the execution of the pipeline including how many times the pipeline was run, when, which components passed or failed, among others.

  The journal was a debug tool but now its user consumable. To accommodate this the format of the journal changed to have more information.

  To save the journal issue the flag --journal myjournal.json.

  See examples/journal/journal.json for an example.

1.9.0

New

• Add option to Mongo sink to allow key overwrite.

  This allows a pipeline to override an entry in the database using the same
  key.

  [sinks.config]
  overwrite = true

• Add type parsing in env source.

  A type can be specified for each environment variable, so that it is parsed
  and collected with the expected datatype. Types available are:

  integer Using Python's int() function.
  float Using Python's float() function.
  string Using Python's str() function.
  auto Using Flowbber's flowbber.utils.types.autocast.
  boolean Using Flowbber's flowbber.utils.types.booleanize.
  iso8601 Using Flowbber's flowbber.utils.iso8601.iso8601_to_datetime.

  Usage:

  [sources.config]
  include = [
      "TESTENV_INT",
  ]
  
  [sources.config.types]
  TESTENV_INT = "integer"

• Pipelines can now be defined in YAML format.

  For example:

  sources:
    - type: timestamp
      id: timestamp
      config:
        epochf: true
        iso8601: true
        strftime: '%Y-%m-%d %H:%M:%S'

Fixes

• Use which genhtml to find executable on lcov_html sink.

  This fixes an issue where the executable could not be found if a custom PATH was used.

1.8.0

New

• The gtest source now supports XML files generated by gtest 1.8.1+.
• New include_files and exclude_files options in many Sinks and Sources.
  See FilterSink Options for more information.
• New compress option added to the archive sink allow to create compressed ZIP archives.
• New extract option added to the JSON source allow to load JSON files from ZIP archives.
• The --extract and --derive-func-data options are now available to use on the LCOV source.

Changes

• Updated schemas to use Cerberus >=1.3.1 definition.

1.7.0

New

• New --dry-run flag allows to parse, load, validate and build a pipeline without executing it.

Changes

• Improved logging when trying to instance a component to help debugging a pipeline that went wrong.
• Improved logging to show a log in higher level when things go bad.

Fixes

• Fix for missing plugin entries in documentation.
• Fix for documentation issue #27.

1.6.0

New

• New LCOV merger aggregator allows to sum multiple LCOV sources.

Fixes

• Fix a bug that ignored rc_overrides when using a file input in LCOV
  source.

1.5.0

Changes

• lcov source no longer accepts directory as configuration.
  New option source superseded it, and allows to specify a directory to
  generate a tracefile or load one already generated.

1.4.0

New

• Refactored Valgrind source to support loading data from Helgrind and DRD tools.
• New "Expander" aggregator that allows to move subdata to top level. This is useful to load data using JSONSource or similar sources and place it in the top level as if it were data from other anonymous sources. Or to replay a pipeline using previously collected data.

1.3.2

New

• Add support for path in InfluxDB sink.

Fixes

• Fixed flake8 issues shown in new version.

1.3.1

Fixes

• Source for Valgrind's memcheck will now always output the stack attribute as a list.

1.3.0

New

• New Config source that allows to add arbitrary data directly from the pipeline definition.
• All plugins now show the example usage in both JSON and TOML.
• Improved documentation for the memcheck source.

Changes

• The Internet speed source plugin is unavailable as the upstream package providing the measurement is currently broken: fopina/pyspeedtest#15

Fixes

• Fix in pytest source that caused a test case with both failure and error to be overridden by the other: pytest-dev/pytest#2228
• Minor fix in memcheck source plugin that caused output that violates the expected schema.