A framework for inspecting data, based on flowcharts from pyqtgraph. plottr's main aim is to allow the user to define custom data processing flows and plotting. A particular use case is data filtering and plotting.
https://plottr.readthedocs.io (work in progress...)
Plottr is installable from pypi with pip install plottr
Plottr requires either the PyQt5 or Pyside2 gui framework.
To install with PyQt5 or Pyside2 backend you can do
pip install plottr[PyQt5] or pip install plottr[Pyside2]
Note that if you have installed pyqt from (Ana)Conda you should not use any of these
targets but do pip install plottr or install Plottr from conda forge:
conda config --add channels conda-forge
conda config --set channel_priority strict
conda install plottr
To install from source: clone the repo, and install using pip install -e .
You can use the inspectr tool to get a simple overview over QCoDeS database
files, and plot datasets contained in the database.
To use: run plottr-inspectr [--dbpath <path to your .db file>]
As an alternative from the root of the cloned plottr repository, run python apps/inspectr.py [--dbpath <path to your .db file>]
For basic instructions, check out the Notebook Live plotting qcodes data under /doc/examples.
Note: this package is not compatible with the original plottr tool.
You might want to install freshly if you still use the old version.
- python >= 3.8
- the usual: numpy, mpl, ...
- pandas >= 0.22
- xarray
- pyqtgraph >= 0.12.1
See the plottr GitHub Releases page for versions and their changelogs
- added a generalized way for fitting data (or similar procedures that generate analysis parameters from datasets) (#293)
- ... and more, see here https://github.com/toolsforexperiments/plottr/releases/tag/v0.10.0
- Fixed a crash bug with python 3.10 (#257)
- Add new analysis node for creating histogram of data by @wpfff in #241
- bugfix: prevent flushing when we are updating data. by @wpfff in #243
- fix type checking with mypy 920 by @jenshnielsen in #248
- Bump minimum version of pyqtgraph by @jenshnielsen in #247
- Test with python 3.10 and mark as supported (#238)
- add missing init file to config dir (#236)
- Use a regular install in tests (#237)
- Inspectr: tag a run with a star (⭐) or cross (❌) icon, filter by those, also show dataset metadata next to parameters and snapshot (#229)
- Improvements to monitr: more stability in adding data to ddh5, better performance by making data loading multithreaded and running reach plot window in a separate process (#219)
- Added pyqtgraph backend for plotting that can be used instead of matplotlib (Example for how to select can be found in test/apps/autoplot_app.py) (#215, #218)
- Fix/invaliddata: small fixes when data contains a lot of invalid entries (#226)
- Fix in shape recognition when adding data (#220)
- Add minimal versions to dependencies (#201)
- Make the .gitignore proper (#73)
- add dependabot (#208)
- Fix typechecking with mypy 0.9xx (#207)
- clarify install instructions wrt qt and mention conda forge (#202)
- refactoring the plotting system (#166)
- Add version log message to main
__init__(#175)
- Fix crop if less than one row is not nan (#198)
- Fix rgba error (#199)
- Allow empty dataset if datadict is none (#195)
- Modernize setup files (#194)
- packaging cleanups (#177)
- upgrade versioneer to 0.19 (#176)
- Add copy content features to inspectr and autoplot windows, specifically
- a new Copy pop up menu for copying content of cells in inspectr
- a new Copy metadata button in plot window for copying info about the dataset to clipboard
- remove redundant information between the optional "info" box on the plot and the plot title
- Drop support for Python 3.6 and support type-checking with qcodes 0.21.0
- Fix type-checking with numpy 1.20
- Fixed y-axis to not show axis-label if more than one plot is selected in 1D single-plot show.
- Workaround for bug with pyqt installed via conda that would result in blank icons.
- Add experimental support for using Pyside2 as an alternative to PyQt5
- Entry points for inspectr (plottr-inspectr) and autoplot (plottr-autoplot-ddh5)
- LICENSE file has been added (no change to license of the code)
- setup.py has been tweeked to ensure that sdist and bdist_wheel packages are generated correctly
- Usable version of hdf5 file support. can use the apps/monitr.py script to launch a tool that allows easy (live) plotting.
- there were several issues/annoyances that caused (minor) issues with grid handling and plotting in rare-ish circumstances. Most of them should be fixed now.
- A few docs and examples, mostly about grid usage.
- the main repo has now moved to to a different organization: https://github.com/toolsforexperiments/plottr the old repo (https://github.com/data-plottr/plottr) will be kept in sync for a while, but deleted eventually.