Arviz

New features

- Added `to_dataframe` method to InferenceData ([1395](https://github.com/arviz-devs/arviz/pull/1395))
- Added `__getitem__` magic to InferenceData ([1395](https://github.com/arviz-devs/arviz/pull/1395))
- Added group argument to summary ([1408](https://github.com/arviz-devs/arviz/pull/1408))
- Add `ref_line`, `bar`, `vlines` and `marker_vlines` kwargs to `plot_rank` ([1419](https://github.com/arviz-devs/arviz/pull/1419))
- Add observed argument to (un)plot observed data in `plot_ppc` ([1422](https://github.com/arviz-devs/arviz/pull/1422))
- Add support for named dims and coordinates with multivariate observations ([1429](https://github.com/arviz-devs/arviz/pull/1429))
- Add support for discrete variables in rank plots ([1433](https://github.com/arviz-devs/arviz/pull/1433)) and
`loo_pit` ([1500](https://github.com/arviz-devs/arviz/pull/1500))
- Add `skipna` argument to `plot_posterior` ([1432](https://github.com/arviz-devs/arviz/pull/1432))
- Make stacking the default method to compute weights in `compare` ([1438](https://github.com/arviz-devs/arviz/pull/1438))
- Add `copy()` method to `InferenceData` class. ([1501](https://github.com/arviz-devs/arviz/pull/1501)).

Maintenance and fixes

- prevent wrapping group names in InferenceData repr_html ([1407](https://github.com/arviz-devs/arviz/pull/1407))
- Updated CmdStanPy interface ([1409](https://github.com/arviz-devs/arviz/pull/1409))
- Remove left out warning about default IC scale in `compare` ([1412](https://github.com/arviz-devs/arviz/pull/1412))
- Fixed a typo found in an error message raised in `distplot.py` ([1414](https://github.com/arviz-devs/arviz/pull/1414))
- Fix typo in `loo_pit` extraction of log likelihood ([1418](https://github.com/arviz-devs/arviz/pull/1418))
- Have `from_pystan` store attrs as strings to allow netCDF storage ([1417](https://github.com/arviz-devs/arviz/pull/1417))
- Remove ticks and spines in `plot_violin` ([1426 ](https://github.com/arviz-devs/arviz/pull/1426))
- Use circular KDE function and fix tick labels in circular `plot_trace` ([1428](https://github.com/arviz-devs/arviz/pull/1428))
- Fix `pair_plot` for mixed discrete and continuous variables ([1434](https://github.com/arviz-devs/arviz/pull/1434))
- Fix in-sample deviance in `plot_compare` ([1435](https://github.com/arviz-devs/arviz/pull/1435))
- Fix computation of weights in compare ([1438](https://github.com/arviz-devs/arviz/pull/1438))
- Avoid repeated warning in summary ([1442](https://github.com/arviz-devs/arviz/pull/1442))
- Fix hdi failure with boolean array ([1444](https://github.com/arviz-devs/arviz/pull/1444))
- Automatically get the current axes instance for `plt_kde`, `plot_dist` and `plot_hdi` ([1452](https://github.com/arviz-devs/arviz/pull/1452))
- Add grid argument to manually specify the number of rows and columns ([1459](https://github.com/arviz-devs/arviz/pull/1459))
- Switch to `compact=True` by default in our plots ([1468](https://github.com/arviz-devs/arviz/issues/1468))
- `plot_elpd`, avoid modifying the input dict ([1477](https://github.com/arviz-devs/arviz/issues/1477))
- Do not plot divergences in `plot_trace` when `kind=rank_vlines` or `kind=rank_bars` ([1476](https://github.com/arviz-devs/arviz/issues/1476))
- Allow ignoring `observed` argument of `pymc3.DensityDist` in `from_pymc3` ([1495](https://github.com/arviz-devs/arviz/pull/1495))
- Make `from_pymc3` compatible with theano-pymc 1.1.0 ([1495](https://github.com/arviz-devs/arviz/pull/1495))
- Improve typing hints ([1491](https://github.com/arviz-devs/arviz/pull/1491), ([1492](https://github.com/arviz-devs/arviz/pull/1492),
([1493](https://github.com/arviz-devs/arviz/pull/1493), ([1494](https://github.com/arviz-devs/arviz/pull/1494) and
([1497](https://github.com/arviz-devs/arviz/pull/1497))

Deprecation

- `plot_khat` deprecate `annotate` argument in favor of `threshold`. The new argument accepts floats ([1478](https://github.com/arviz-devs/arviz/issues/1478))

Documentation

- Reorganize documentation and change sphinx theme ([1406](https://github.com/arviz-devs/arviz/pull/1406))
- Switch to [MyST](https://myst-parser.readthedocs.io/en/latest/) and [MyST-NB](https://myst-nb.readthedocs.io/en/latest/index.html)
for markdown/notebook parsing in docs ([1406](https://github.com/arviz-devs/arviz/pull/1406))
- Incorporated `input_core_dims` in `hdi` and `plot_hdi` docstrings ([1410](https://github.com/arviz-devs/arviz/pull/1410))
- Add documentation pages about experimental `SamplingWrapper`s usage ([1373](https://github.com/arviz-devs/arviz/pull/1373))
- Show example titles in gallery page ([1484](https://github.com/arviz-devs/arviz/pull/1484))
- Add `sample_stats` naming convention to the InferenceData schema ([1063](https://github.com/arviz-devs/arviz/pull/1063))
- Extend api documentation about `InferenceData` methods ([1338](https://github.com/arviz-devs/arviz/pull/1338))

Experimental

- Modified `SamplingWrapper` base API ([1373](https://github.com/arviz-devs/arviz/pull/1373))

New features

- Added InferenceData dataset containing circular variables ([1265](https://github.com/arviz-devs/arviz/pull/1265))
- Added `is_circular` argument to `plot_dist` and `plot_kde` allowing for a circular histogram (Matplotlib, Bokeh) or 1D KDE plot (Matplotlib). ([1266](https://github.com/arviz-devs/arviz/pull/1266))
- Added `to_dict` method for InferenceData object ([1223](https://github.com/arviz-devs/arviz/pull/1223))
- Added `circ_var_names` argument to `plot_trace` allowing for circular traceplot (Matplotlib) ([1336](https://github.com/arviz-devs/arviz/pull/1336))
- Ridgeplot is hdi aware. By default displays truncated densities at the specified `hdi_prop` level ([1348](https://github.com/arviz-devs/arviz/pull/1348))
- Added `plot_separation` ([1359](https://github.com/arviz-devs/arviz/pull/1359))
- Extended methods from `xr.Dataset` to `InferenceData` ([1254](https://github.com/arviz-devs/arviz/pull/1254))
- Add `extend` and `add_groups` to `InferenceData` ([1300](https://github.com/arviz-devs/arviz/pull/1300) and [1386](https://github.com/arviz-devs/arviz/pull/1386))
- Added `__iter__` method (`.items`) for InferenceData ([1356](https://github.com/arviz-devs/arviz/pull/1356))
- Add support for discrete variables in `plot_bpv` ([1379](https://github.com/arviz-devs/arviz/pull/1379))

Maintenance and fixes

- Automatic conversion of list/tuple to numpy array in distplot ([1277](https://github.com/arviz-devs/arviz/pull/1277))
- `plot_posterior` fix overlap of hdi and rope ([1263](https://github.com/arviz-devs/arviz/pull/1263))
- `plot_dist` bins argument error fixed ([1306](https://github.com/arviz-devs/arviz/pull/1306))
- Improve handling of circular variables in `az.summary` ([1313](https://github.com/arviz-devs/arviz/pull/1313))
- Removed change of default warning in `ELPDData` string representation ([1321](https://github.com/arviz-devs/arviz/pull/1321))
- Update `radon` example dataset to current InferenceData schema specification ([1320](https://github.com/arviz-devs/arviz/pull/1320))
- Update `from_cmdstan` functionality and add warmup groups ([1330](https://github.com/arviz-devs/arviz/pull/1330) and [1351](https://github.com/arviz-devs/arviz/pull/1351))
- Restructure plotting code to be compatible with mpl>=3.3 ([1312](https://github.com/arviz-devs/arviz/pull/1312) and [1352](https://github.com/arviz-devs/arviz/pull/1352))
- Replaced `_fast_kde()` with `kde()` which now also supports circular variables via the argument `circular` ([1284](https://github.com/arviz-devs/arviz/pull/1284)).
- Increased `from_pystan` attrs information content ([1353](https://github.com/arviz-devs/arviz/pull/1353))
- Allow `plot_trace` to return and accept axes ([1361](https://github.com/arviz-devs/arviz/pull/1361))
- Update diagnostics to be on par with posterior package ([1366](https://github.com/arviz-devs/arviz/pull/1366))
- Use method="average" in `scipy.stats.rankdata` ([1380](https://github.com/arviz-devs/arviz/pull/1380))
- Add more `plot_parallel` examples ([1380](https://github.com/arviz-devs/arviz/pull/1380))
- Bump minimum xarray version to 0.16.1 ([1389](https://github.com/arviz-devs/arviz/pull/1389)
- Fix multi rope for `plot_forest` ([1390](https://github.com/arviz-devs/arviz/pull/1390))
- Bump minimum xarray version to 0.16.1 ([1389](https://github.com/arviz-devs/arviz/pull/1389))
- `from_dict` will now store warmup groups even with the main group missing ([1386](https://github.com/arviz-devs/arviz/pull/1386))
- increase robustness for repr_html handling ([1392](https://github.com/arviz-devs/arviz/pull/1392))

New features

- loo-pit plot. The kde is computed over the data interval (this could be shorter than [0, 1]). The HDI is computed analytically ([1215](https://github.com/arviz-devs/arviz/pull/1215))
- Added `html_repr` of InferenceData objects for jupyter notebooks. ([1217](https://github.com/arviz-devs/arviz/pull/1217))
- Added support for PyJAGS via the function `from_pyjags`. ([1219](https://github.com/arviz-devs/arviz/pull/1219) and [1245](https://github.com/arviz-devs/arviz/pull/1245))
- `from_pymc3` can now retrieve `coords` and `dims` from model context ([1228](https://github.com/arviz-devs/arviz/pull/1228), [1240](https://github.com/arviz-devs/arviz/pull/1240) and [1249](https://github.com/arviz-devs/arviz/pull/1249))
- `plot_trace` now supports multiple aesthetics to identify chain and variable
shape and support matplotlib aliases ([1253](https://github.com/arviz-devs/arviz/pull/1253))
- `plot_hdi` can now take already computed HDI values ([1241](https://github.com/arviz-devs/arviz/pull/1241))
- `plot_bpv`. A new plot for Bayesian p-values ([1222](https://github.com/arviz-devs/arviz/pull/1222))

Maintenance and fixes

- Include data from `MultiObservedRV` to `observed_data` when using
`from_pymc3` ([1098](https://github.com/arviz-devs/arviz/pull/1098))
- Added a note on `plot_pair` when trying to use `plot_kde` on `InferenceData`
objects. ([1218](https://github.com/arviz-devs/arviz/pull/1218))
- Added `log_likelihood` argument to `from_pyro` and a warning if log likelihood cannot be obtained ([1227](https://github.com/arviz-devs/arviz/pull/1227))
- Skip tests on matplotlib animations if ffmpeg is not installed ([1227](https://github.com/arviz-devs/arviz/pull/1227))
- Fix hpd bug where arguments were being ignored ([1236](https://github.com/arviz-devs/arviz/pull/1236))
- Remove false positive warning in `plot_hdi` and fixed matplotlib axes generation ([1241](https://github.com/arviz-devs/arviz/pull/1241))
- Change the default `zorder` of scatter points from `0` to `0.6` in `plot_pair` ([1246](https://github.com/arviz-devs/arviz/pull/1246))
- Update `get_bins` for numpy 1.19 compatibility ([1256](https://github.com/arviz-devs/arviz/pull/1256))
- Fixes to `rug`, `divergences` arguments in `plot_trace` ([1253](https://github.com/arviz-devs/arviz/pull/1253))

Deprecation

- Using `from_pymc3` without a model context available now raises a
`FutureWarning` and will be deprecated in a future version ([1227](https://github.com/arviz-devs/arviz/pull/1227))
- In `plot_trace`, `chain_prop` and `compact_prop` as tuples will now raise a
`FutureWarning` ([1253](https://github.com/arviz-devs/arviz/pull/1253))
- `hdi` with 2d data raises a FutureWarning ([1241](https://github.com/arviz-devs/arviz/pull/1241))

Documentation

- A section has been added to the documentation at InferenceDataCookbook.ipynb illustrating the use of ArviZ in conjunction with PyJAGS. ([1219](https://github.com/arviz-devs/arviz/pull/1219) and [1245](https://github.com/arviz-devs/arviz/pull/1245))
- Fixed inconsistent capitalization in `plot_hdi` docstring ([1221](https://github.com/arviz-devs/arviz/pull/1221))
- Fixed and extended `InferenceData.map` docs ([1255](https://github.com/arviz-devs/arviz/pull/1255))

Maintenance and fixes

- Restructured internals of `from_pymc3` to handle old pymc3 releases and
sliced traces and to provide useful warnings ([1211](https://github.com/arviz-devs/arviz/pull/1211))

Maintenance and fixes

- Fixed bug in `from_pymc3` for sliced `pymc3.MultiTrace` input ([1209](https://github.com/arviz-devs/arviz/pull/1209))

Maintenance and fixes

- Fixed bug in `from_pymc3` when used with PyMC3<3.9 ([1203](https://github.com/arviz-devs/arviz/pull/1203))
- Fixed enforcement of rcParam `plot.max_subplots` in `plot_trace` and
`plot_pair` ([1205](https://github.com/arviz-devs/arviz/pull/1205))
- Removed extra subplot row and column in in `plot_pair` with `marginal=True` ([1205](https://github.com/arviz-devs/arviz/pull/1205))
- Added latest PyMC3 release to CI in addition to using GitHub default branch ([1207](https://github.com/arviz-devs/arviz/pull/1207))

Documentation

- Use `dev` as version indicator in online documentation ([1204](https://github.com/arviz-devs/arviz/pull/1204))