Skip to content

Latest commit

 

History

History
347 lines (189 loc) · 6.69 KB

CHANGELOG.rst

File metadata and controls

347 lines (189 loc) · 6.69 KB
Raw

Change Log

v0.8.2

Added:

None

  • Add installation scripts for closed source packages like diffpy.pdfgetx

Changed:

  • Use multiprocess in pathos instead of multiprocessing to run the tests for servers

Fixed:

  • Update the code to fit the latest version of databroker and ophyd

v0.8.1

Fixed:

  • xpdan.vend.callbacks.core.Retrieve does the right thing when there is no filled key

v0.8.0

Added:

  • xpdan.startup.instensity_server for tracking positions in 1D patterns
  • peak position and width server
  • intensity server to setup.py
  • stage_blacklist to analysis server, allowing for speedups by not passing large arrays back to the proxy.
  • xpdan.vend.callbacks.mpl_plotting.SavePlots callback for saving plots when a scan is finished
  • save_folder kwarg for viz server. If provided save figures

Fixed:

  • analysis_server kwargs are properly passed
  • analysis server reuses diffraction pipeline (getting rid of memory issues)

v0.7.0

Added:

  • Pull detector name from descriptor if possible
  • Analysis system for radiograph data processing
  • Routers for diffraction and radiograph analysis
  • zscore kwarg to analysis pipeline

v0.6.1

Changed:

  • mayavi callback is accessed via a shim in case it doesn't exist

v0.6.0

Added:

  • Add inbound_proxy_address so we can access the publisher at runtime
  • Add qoi_server CLI and module for computing various quantities of interest
  • Added amorphsivity to QOIs
  • Support for 3D pencil beam reconstructions
  • Support for Registry backed full field reconstructions
  • ctPDF_example.py example for pencil beam

Changed:

  • Use image_fields for loading the images
  • tomo_example.py runs on a registry backed detector

Fixed:

  • Pencil beam no longer tries to run independent var reconstructions

v0.5.2

Fixed:

  • Image streams specifically listen to the primary event stream
  • AlignEventStream with the raw data listens to the primary event stream

v0.5.1

Fixed:

  • Fix mask overlay order so it isn't delayed

v0.5.0

Added:

  • Units for most reduced data streams (mostly so waterfall can plot them)
  • Added plot_graph to analysis server so we can visualize the graph

Changed:

  • Use dark_frame=True to determine if a scan is a dark scan. This means that scans without darks can now be processed, although they will use the most recenlty cached dark.
  • raw_stripped is run first, so all the aligned data can be emitted ASAP. This means that data can be saved as soon as it is created, rather than waiting for the last data to be processed.
  • Simplified the plotting via the new LiveWaterfall callback.
  • Support original_start_uid system
  • use raw start document time for file naming

v0.4.1

v0.4.0

Added:

  • analysis server for running analysis in an external process without saving or visualization
  • DB server for saving data into databrokers
  • ExportCallback for copying data for databroker export
  • Retrieve callback for accessing data from resources and datums
  • ReturnCallback a callback which always returns name document pairs
  • RunRouter which allows us to make pipelines on the fly
  • Server for handling saving of data
  • Tests for the servers
  • All the callbacks from bluesky

Changed:

  • Chunk the analysis pipelines into pieces so they can be composed
  • StartStopCallback also reports the analysis time for diagnostics
  • Multiple detectors are supported for the main processing pipeline note that this does not support multiple detectors at the same time.
  • StartStopCallback now prints event analysis times
  • Moved file saving print statements closer to saving action
  • Moved hinted_fields to the xpdan.vend.callbacks.core
  • BestEffortCallback has a teardown kwarg for managing figures at the end of a run, preventing us from having too many windows open.

Fixed:

  • Removed sanitize_np from zmq system.

Authors:

  • Christopher J. Wright

v0.3.6

Fixed:

  • Both the foreground dark and foreground light are properly pulled based off their stream names

v0.3.5

Fixed:

  • Restore calib saving
  • Make certain that files are saved before additional analysis is performed. This should make the pipeline more robust to analysis failures near the end of the pipeline.
  • xpdan.callbacks.StopStartCallback to be no-stop doc tolerant
  • conftest.py now properly reports a bt_uid
  • xpdan.db_utils.query_dark now always returns a list

v0.3.4

Added:

  • save_tiff pipeline for simple save tiff

v0.3.3

Fixed:

  • Splay out args for calibration saving properly
  • Pull calibration information from dSpacings

v0.3.2

Fixed:

  • Cast to numpy float32 on way into pipeline

v0.3.1

Added:

  • Add print statement for server startup

Changed:

  • metadata now stored into its own folder

Fixed:

  • Files saved in sub dir of tiff_base
  • Filter out None produced by Filler on datum/resource docs

v0.3.0

Added:

  • Filler and clear_combine_latest to pipeline_utils
  • Tests for main pipeline
  • Support for QOI plotting
  • start_analysis function for starting the RemoteDispatcher and pushing data through the pipeline live

Changed:

  • db_utils query functions signatures are now reversed (docs first, db second)
  • Simplified render_and_clean
  • MainCallback analysis_stage syntax in line with pipeline
  • conftest databroker now reports raw dicts
  • Separated saving, visualization, and QOI pipeline chunks
  • Darks and Backgrounds are zeroed out at start of scan in case scan has no dark or background.

Removed:

  • xpdan.tools is now in xpdtools.tools
  • shim module is now in xpdtools

Fixed:

  • Analysis pipeline now runs on xpdtools pipeline

v0.2.3rc

Changed:

  • xpdAcq now outsources ``glbl` configuration management to xpdConf
  • Use conda-forge xpdconf for CI builds

Removed:

  • load_configuration (which is now in xpdConf)

v0.2.2

v0.2.1

Added:

  • Requirements folder
  • push_tag to rever
  • Flexible folder tag
  • Add rever changelog activity
  • Speed up masking via median based sigma clipping
  • Z score visualization to callback pipeline