### Install ArviZ from Source

Source: https://github.com/arviz-devs/arviz/blob/main/README.md

Clone the ArviZ repository and install it locally using setup.py. This method is useful for developers or for installing a specific commit.

```bash
git clone https://github.com/arviz-devs/arviz.git
cd arviz
python setup.py install
```

--------------------------------

### Install and Test ArviZ Locally

Source: https://github.com/arviz-devs/arviz/wiki/ArviZ-Release-Checklist

Install ArviZ in development mode and run the test suite. Ensure this is done after updating the version number.

```bash
python setup.py develop
pytest arviz/tests
```

--------------------------------

### Install ArviZ with Preview Features

Source: https://github.com/arviz-devs/arviz/blob/main/README.md

Use this command to install the latest stable version of ArviZ along with preview features. This is useful for accessing the newest functionalities.

```bash
pip install "arviz[preview]"
```

--------------------------------

### Install Development Environment

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Install tox and the ArviZ local package in editable mode to set up your development environment.

```bash
pip install tox
pip install -e .
```

--------------------------------

### Install ArviZ with optional backends

Source: https://context7.com/arviz-devs/arviz/llms.txt

Install ArviZ using pip, specifying optional dependencies for plotting backends (Matplotlib, Bokeh, Plotly) and I/O support (NetCDF, Zarr). Conda installation is also available.

```bash
pip install "arviz[matplotlib,h5netcdf]"

```

```bash
pip install "arviz[zarr,bokeh]"

```

```bash
pip install "arviz[plotly]"

```

```bash
conda install -c conda-forge arviz arviz-plots

```

--------------------------------

### Install ArviZ with Optional Dependencies

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/getting_started.md

Use this command to install ArviZ with specific optional features like I/O and plotting capabilities. Choose the features you need based on the available options.

```bash
pip install "arviz[<I/O>, <plotting>]"
```

```bash
pip install "arviz[zarr, matplotlib]"
```

```bash
pip install "arviz[h5netcdf, plotly, bokeh]"
```

--------------------------------

### Verify PyPI Installation

Source: https://github.com/arviz-devs/arviz/wiki/ArviZ-Release-Checklist

After releasing to PyPI, install the new version and verify that ArviZ can be imported. This confirms the release was successful.

```python
python -c 'from arviz import *'
```

--------------------------------

### Load Built-in ArviZ Example Datasets

Source: https://context7.com/arviz-devs/arviz/llms.txt

Use `az.load_arviz_data` to load example datasets provided with ArviZ. This is useful for testing, learning, and reproducing documentation examples. The function returns an `InferenceData` object as a `DataTree`.

```python
import arviz as az

# List available datasets (see docs for full list)
# Includes: "centered_eight", "non_centered_eight", "radon", "rugby", etc.

dt = az.load_arviz_data("centered_eight")
print(list(dt.children))
# ['posterior', 'posterior_predictive', 'log_likelihood',
#  'sample_stats', 'prior', 'prior_predictive', 'observed_data', 'constant_data']

dt_nc = az.load_arviz_data("non_centered_eight")
print(dt_nc["posterior"].dataset)
# <xarray.Dataset>
# Dimensions:  (chain: 4, draw: 500, school: 8)
# Data variables:
#     mu       (chain, draw) float64  ...
#     theta_t  (chain, draw, school) float64  ...
#     tau      (chain, draw) float64  ...
#     theta    (chain, draw, school) float64  ...
```

--------------------------------

### Verify ArviZ Installation

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/getting_started.md

Run this Python code to check if ArviZ is installed correctly and to display its version along with the versions of its constituent libraries. This helps in troubleshooting installation issues.

```python
import arviz as az
print(az.info)
```

--------------------------------

### Build and Upload ArviZ to PyPI

Source: https://github.com/arviz-devs/arviz/wiki/ArviZ-Release-Checklist

Create source and wheel distributions and upload them to PyPI using Twine. Ensure Twine is installed and you have the necessary credentials.

```bash
python setup.py sdist bdist_wheel
twine upload dist/*
```

--------------------------------

### Install ArviZ Development Version from GitHub

Source: https://github.com/arviz-devs/arviz/blob/main/README.md

Install the latest development version of ArviZ directly from its GitHub repository using pip. This is suitable for users who want to test the most recent code.

```bash
pip install git+git://github.com/arviz-devs/arviz.git
```

--------------------------------

### az.load_arviz_data

Source: https://context7.com/arviz-devs/arviz/llms.txt

Loads one of ArviZ's bundled example datasets as a `DataTree`. This function is useful for testing, learning, and reproducing examples from the documentation.

```APIDOC
## `az.load_arviz_data` — Load built-in example datasets

Loads one of ArviZ's bundled example datasets as a `DataTree`. Useful for testing, learning, and reproducing examples from the documentation.

```python
import arviz as az

# List available datasets (see docs for full list)
# Includes: "centered_eight", "non_centered_eight", "radon", "rugby", etc.

dt = az.load_arviz_data("centered_eight")
print(list(dt.children))
# ['posterior', 'posterior_predictive', 'log_likelihood',
#  'sample_stats', 'prior', 'prior_predictive', 'observed_data', 'constant_data']

dt_nc = az.load_arviz_data("non_centered_eight")
print(dt_nc["posterior"].dataset)
# <xarray.Dataset>
# Dimensions:  (chain: 4, draw: 500, school: 8)
# Data variables:
#     mu       (chain, draw) float64 ...
#     theta_t  (chain, draw, school) float64 ...
#     tau      (chain, draw) float64 ...
#     theta    (chain, draw, school) float64 ...
```
```

--------------------------------

### See Also Section Example

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/docstrings.md

Illustrates how to format the 'See Also' section in a docstring, including function names and brief descriptions. Sphinx automatically creates links for listed ArviZ objects.

```python
    See Also
    --------
    arviz_stats.hdi : Calculate highest density interval (HDI) of array for given probability.
    arviz_plots.plot_ppc_dist : plot for posterior/prior predictive checks.
```

--------------------------------

### Verify ArviZ installation and component versions

Source: https://context7.com/arviz-devs/arviz/llms.txt

Use `az.info` to check the ArviZ version and confirm the availability of its sub-libraries (`arviz_base`, `arviz_stats`, `arviz_plots`). This is useful for troubleshooting import issues. Detailed logging can be enabled for further diagnostics.

```python
import arviz as az

print(az.info)
# Status information for ArviZ 1.1.0
#
# arviz_base 1.1.0 available, exposing its functions as part of the `arviz` namespace
# arviz_stats 1.1.0 available, exposing its functions as part of the `arviz` namespace
# arviz_plots 1.1.0 available, exposing its functions as part of the `arviz` namespace

# Enable detailed logging if a library fails to import
import logging
logging.basicConfig(level=logging.INFO)
import arviz as az

```

--------------------------------

### Install ArviZ and ArviZ-Plots via Conda

Source: https://github.com/arviz-devs/arviz/blob/main/README.md

Install ArviZ and its plotting companion package using conda-forge. This command ensures both core ArviZ functionality and plotting tools are set up.

```bash
conda install -c conda-forge arviz arviz-plots
```

--------------------------------

### Setting computational backend to 'base'

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Configure ArviZ to use the 'base' computational backend for statistics functions by setting `az.rcParams["stats.module"]`. This example also times the execution of the `histogram` method.

```python
dt = az.load_arviz_data("radon")
az.rcParams["stats.module"] = "base"
%timeit dt.azstats.histogram(dim="draw")
```

--------------------------------

### Setting computational backend to 'numba'

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Configure ArviZ to use the 'numba' computational backend for statistics functions by setting `az.rcParams["stats.module"]`. This example also times the execution of the `histogram` method.

```python
az.rcParams["stats.module"] = "numba"
%timeit dt.azstats.histogram(dim="draw")
```

--------------------------------

### Suggesting Matplotlib References with sphobjinv

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/syntax_guide.md

Example of using the sphobjinv command-line tool to suggest possible references for 'axes.plot' from Matplotlib documentation.

```bash
$ sphobjinv suggest -t 90 -u https://matplotlib.org/objects.inv "axes.plot"

Remote inventory found.

:py:method:`matplotlib.axes.Axes.plot`
:py:method:`matplotlib.axes.Axes.plot_date`
:std:doc:`api/_as_gen/matplotlib.axes.Axes.plot`
:std:doc:`api/_as_gen/matplotlib.axes.Axes.plot_date`
```

--------------------------------

### az.info — Verify installation and component versions

Source: https://context7.com/arviz-devs/arviz/llms.txt

A string attribute that reports the ArviZ version and the availability of each sub-library (`arviz_base`, `arviz_stats`, `arviz_plots`). Useful for troubleshooting import issues.

```APIDOC
## az.info

### Description
Verifies ArviZ installation and component versions.

### Usage
```python
import arviz as az

print(az.info)
```

### Example Output
```
Status information for ArviZ 1.1.0

arviz_base 1.1.0 available, exposing its functions as part of the `arviz` namespace
arviz_stats 1.1.0 available, exposing its functions as part of the `arviz` namespace
arviz_plots 1.1.0 available, exposing its functions as part of the `arviz` namespace
```
```

--------------------------------

### Legacy InferenceData Migration Warning

Source: https://context7.com/arviz-devs/arviz/llms.txt

Shows how ArviZ raises a `MigrationWarning` when legacy `InferenceData` is accessed, directing users to the migration guide and returning `xarray.DataTree`.

```python
import arviz as az
import warnings

# Accessing legacy InferenceData triggers a warning and returns DataTree
with warnings.catch_warnings(record=True) as w:
    warnings.simplefilter("always")
    from arviz import InferenceData
    assert issubclass(w[0].category, az.MigrationWarning)
    assert "migration_guide" in str(w[0].message)

# InferenceData is now DataTree
from xarray import DataTree
assert InferenceData is DataTree

# Migration guide: https://python.arviz.org/en/latest/user_guide/migration_guide.html
```

--------------------------------

### Load Inference Data from NetCDF or Zarr

Source: https://context7.com/arviz-devs/arviz/llms.txt

Use `az.from_netcdf` to load data from NetCDF files and `az.from_zarr` for Zarr stores. These functions are compatible with ArviZ's `InferenceData` objects and can also load data saved from legacy formats. The example demonstrates saving and reloading data for a round-trip check.

```python
import arviz as az

# Load a NetCDF file (engine auto-detected or explicit)
dt = az.from_netcdf("inference_results.nc")
dt = az.from_netcdf("inference_results.nc", engine="h5netcdf")

# Load a Zarr store
dt = az.from_zarr("inference_results.zarr")

# Save and reload round-trip example
dt = az.load_arviz_data("centered_eight")
dt.to_netcdf("example.nc", engine="h5netcdf")
dt_reloaded = az.from_netcdf("example.nc")

print(dt_reloaded.groups)
# ('/', '/posterior', '/posterior_predictive', '/log_likelihood',
#  '/sample_stats', '/prior', '/prior_predictive', '/observed_data', '/constant_data')

# Access a specific group as DataTree or Dataset
posterior_dt = dt_reloaded["posterior"]           # DataTree
posterior_ds = dt_reloaded["posterior"].dataset    # Dataset (view)
theta = dt_reloaded["posterior"]["theta"]          # DataArray (unchanged)
```

--------------------------------

### Compute R-hat Convergence Diagnostic

Source: https://context7.com/arviz-devs/arviz/llms.txt

Compute the rank-normalized R-hat statistic using `az.rhat`. Values near 1.0 suggest chain convergence, while values above 1.01–1.05 indicate potential mixing issues. The example shows how to check for problematic variables.

```python
import arviz as az

dt = az.load_arviz_data("centered_eight")
rhat_result = az.rhat(dt)
print(rhat_result)
# <xarray.DataTree 'posterior'>
# Group: /posterior
#     Data variables:
#         mu     float64  1.001
#         theta  (school) float64  1.002 1.000 ... 1.001 1.001
#         tau    float64  1.003

# Check for problematic variables (rhat > 1.01)
import xarray as xr
rhat_ds = rhat_result["posterior"].dataset
problematic = {
    var: float(rhat_ds[var].max())
    for var in rhat_ds
    if float(rhat_ds[var].max()) > 1.01
}
print(problematic)  # e.g. {'tau': 1.025}
```

--------------------------------

### Atomic Plot Function Example

Source: https://github.com/arviz-devs/arviz/wiki/Plot-hierarchy

Demonstrates the signature and common input parameters for atomic plot functions, which operate on a single axis. These functions typically accept data arrays, sources, axes, labels, and keyword arguments for plot customization.

```python
*arr: one or two {ndarray, str}
source: {xarray.Dataset, pandas.DataFrame,bokeh.ColumnDataSource}; optional
ax: {matplotlib.axis, bokeh.figure}; optional
labels: one or two {str, bokeh.Text}
line_kwargs: {dict}; optional
    line properties
fill_kwargs: {dict}; optional
    area properties
special_kwargs: {dict}; optional
    special plot specific properties
backend: {"matplotlib","bokeh"} 
```

--------------------------------

### Build and View ArviZ Docs

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/sphinx_doc_build.md

Use these commands to build the documentation with Sphinx and then view it in a browser. Re-run the build command after making changes.

```bash
tox -e docs      # build docs with sphinx
tox -e viewdocs  # view docs in browser
```

--------------------------------

### Build Documentation with Tox

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_checklist.md

Use this command to build the project's documentation after incorporating your changes. This ensures that the documentation is up-to-date and correctly formatted.

```shell
tox -e docs
```

--------------------------------

### Live Preview of ArviZ Docs

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/sphinx_doc_build.md

Enable a live preview of the documentation that automatically updates when changes are saved. This avoids the need to manually rebuild after every modification.

```bash
tox -e livedocs
```

--------------------------------

### Get Default Summary - ArviZ

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/label_guide.rst

Generates a summary of the inference data with its default dimension order.

```python
az.summary(experiments)
```

--------------------------------

### Reduce Dimensions with `hdi`

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Use `dim` to specify dimensions to reduce. This example reduces 'school' and 'draw' dimensions using the `hdi` function.

```python
dt.azstats.hdi(dim=["school", "draw"])
```

--------------------------------

### Load data and display summary

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/label_guide.rst

Loads the 'centered_eight' dataset and displays a summary using default labeling.

```python
import arviz as az
schools = az.load_arviz_data("centered_eight")
az.summary(schools)
```

--------------------------------

### List Available Tox Commands

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

List the available development tasks managed by tox for the ArviZ repository.

```bash
tox list -m dev
```

--------------------------------

### Accessing Groups in InferenceData

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Access the groups attribute of a DataTree object to get a list of Unix directory paths representing the groups. This is an alternative to the old InferenceData.groups.

```python
dt.groups
```

--------------------------------

### Accessing Sample Stats Prior with PyStan

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Access the sample statistics from the prior distribution. This is useful for understanding the prior's behavior.

```python
idata_stan.sample_stats_prior
```

--------------------------------

### Clean and Rebuild ArviZ Docs

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/sphinx_doc_build.md

Execute this command to clean all cache and intermediate files, forcing a complete rebuild of the documentation from scratch. This is useful when changes are not reflected correctly by standard build commands.

```bash
tox -e cleandocs
```

--------------------------------

### Check Code Style with Tox

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_checklist.md

Run this command to ensure your code adheres to the project's style guidelines. It's a crucial step before submitting your pull request.

```shell
tox -e check
```

--------------------------------

### Stage and Commit Changes

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Stage your modified files and commit them with a descriptive message to record your changes locally.

```bash
git add modified_files
git commit -m "commit message here"
```

--------------------------------

### Compute Credible Intervals (HDI/ETI)

Source: https://context7.com/arviz-devs/arviz/llms.txt

Compute Highest Density Intervals (HDI) or Equal-Tailed Intervals (ETI) for posterior variables using `az.hdi` or `az.eti`. These functions reduce over specified dimensions and support arbitrary variable subsets. The example shows HDI calculation with a specified probability.

```python
import arviz as az

dt = az.load_arviz_data("centered_eight")

# HDI over chain and draw (default ci_prob=0.89)
hdi_result = az.hdi(dt, hdi_prob=0.94)
print(hdi_result["posterior"].dataset)
# <xarray.Dataset>
# Dimensions:   (ci_bound: 2, school: 8)

```

--------------------------------

### Sample from Prior Distribution in PyStan

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Samples from the defined prior distribution using PyStan's `sampling` method with `Fixed_param` algorithm for prior predictive checks.

```python
linreg_prior_data_dict = {"N": N, "time_since_joined": time_since_joined}
prior = sm_prior.sampling(
    data=linreg_prior_data_dict, iter=150, chains=1, algorithm="Fixed_param", warmup=0
)
```

--------------------------------

### Compute PSIS-LOO-CV with az.loo

Source: https://context7.com/arviz-devs/arviz/llms.txt

Calculate the Pareto-smoothed importance sampling leave-one-out cross-validation (PSIS-LOO-CV) estimate of predictive performance using az.loo. This function returns an ELPDData object containing LOO estimates, standard errors, and Pareto-k diagnostics. Use `pointwise=True` to get pointwise estimates.

```python
import arviz as az

dt = az.load_arviz_data("centered_eight")

loo_result = az.loo(dt, pointwise=True)
print(loo_result)
```

```python
# Check Pareto-k diagnostic (values < 0.7 are good)
print(loo_result.pareto_k)
```

```python
# LOO with R2 metric
loo_r2 = az.loo_r2(dt)
```

--------------------------------

### Subset and Summarize with IdxLabeller - ArviZ

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/label_guide.rst

Generate a summary of a subsetted inference data object using `IdxLabeller`. Ensure positional indexes are used on the corresponding subset for accurate results.

```python
az.summary(schools.isel(school=[2, 5, 7]), labeller=azl.IdxLabeller())
```

--------------------------------

### Compute Monte Carlo Standard Error (MCSE)

Source: https://context7.com/arviz-devs/arviz/llms.txt

Calculate the Monte Carlo Standard Error (MCSE) for posterior mean and quantile estimates using `az.mcse`. MCSE measures the precision of MCMC estimates and is used with ESS to assess sampling quality. The example shows calculation for means and specific quantiles.

```python
import arviz as az

dt = az.load_arviz_data("centered_eight")
mcse_result = az.mcse(dt)
print(mcse_result["posterior"].dataset)

# MCSE for specific quantiles
mcse_quantile = az.mcse(dt, method="quantile", prob=0.05)
```

--------------------------------

### Add Upstream Remote (SSH/HTTPS)

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Add the base ArviZ repository as a remote named 'upstream' to your local clone, using either SSH or HTTPS.

```bash
cd <arviz repo>
git remote add upstream git@github.com:arviz-devs/<arviz repo>.git
```

```bash
cd <arviz repo>
git remote add upstream https://github.com/arviz-devs/<arviz repo>.git
```

--------------------------------

### Import ArviZ

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Imports the ArviZ library under the conventional alias 'az'.

```python
import arviz as az
```

--------------------------------

### Sample from Posterior Distribution in PyStan

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Samples from the posterior distribution of the linear regression model using PyStan's `sampling` method with specified data and MCMC parameters.

```python
linreg_data_dict = {
    "N": N,
    "slack_comments": slack_comments,
    "github_commits": github_commits,
    "time_since_joined": time_since_joined,
    "N_pred": N_pred,
    "time_since_joined_pred": candidate_devs_time,
}
posterior = sm.sampling(data=linreg_data_dict, iter=200, chains=4)
```

--------------------------------

### Access Sample Stats Data in PyStan

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Retrieve the sample statistics dataset from a PyStan InferenceData object. This includes metrics like acceptance statistics, step size, and tree depth.

```python
idata_stan.sample_stats
```

--------------------------------

### Import Libraries for PyStan and ArviZ

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Imports necessary libraries for data manipulation and ArviZ visualization. Sets display style for xarray.

```python
import arviz as az
import pystan
import pandas as pd
import numpy as np
import xarray

xarray.set_options(display_style="html");
```

--------------------------------

### Sync Local Repository with Upstream Main

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Sync your local main branch with the upstream repository's main branch before creating a new feature branch.

```bash
git checkout main
git fetch upstream
git rebase upstream/main
```

--------------------------------

### Accessing Prior Predictive Samples with PyStan

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Retrieve prior predictive samples generated using PyStan. These samples represent data generated from the prior distributions.

```python
idata_stan.prior_predictive
```

--------------------------------

### Use IdxLabeller for Positional Indexes - ArviZ

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/label_guide.rst

Utilize `IdxLabeller` to display positional indexes instead of coordinate values in the summary. This is helpful for understanding data structure.

```python
az.summary(schools, labeller=azl.IdxLabeller())
```

--------------------------------

### Configuring Sample Dimensions for Data Conversion

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Set the 'data.sample-dims' rcParam to 'sample' to automatically handle data without a chain dimension during conversion. This avoids the need for manual dimension addition.

```python
az.rcParams["data.sample_dims"] = "sample"
```

--------------------------------

### Sync and Push Changes

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Sync your local branch with upstream changes and then push your feature branch to your GitHub fork.

```bash
git fetch upstream
git rebase upstream/main
git push -u origin my-feature
```

--------------------------------

### Model Comparison Visualization

Source: https://context7.com/arviz-devs/arviz/llms.txt

Visualizes the output of `az.compare()` as a dot plot with error bars. This facilitates ranking models and assessing the significance of differences between them.

```python
import arviz as az

dt_c = az.load_arviz_data("centered_eight")
dt_nc = az.load_arviz_data("non_centered_eight")

comparison = az.compare({"centered": dt_c, "non_centered": dt_nc})
az.plot_compare(comparison)
```

--------------------------------

### Clone ArviZ Fork (SSH/HTTPS)

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Clone your forked ArviZ repository to your local machine using either SSH or HTTPS.

```bash
git clone git@github.com:<your GitHub handle>/<arviz repo>.git
```

```bash
git clone https://github.com/<your GitHub handle>/<arviz repo>.git
```

--------------------------------

### Import Libraries for PyMC3 and ArviZ

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyMC3_schema_example.ipynb

Imports necessary libraries including ArviZ, PyMC3, pandas, numpy, and xarray. Sets xarray display style to HTML for better rendering.

```python
import arviz as az
import pymc3 as pm
import pandas as pd
import numpy as np
import xarray

xarray.set_options(display_style="html");
```

--------------------------------

### az.from_netcdf and az.from_zarr

Source: https://context7.com/arviz-devs/arviz/llms.txt

Load inference data from NetCDF or Zarr formats. `az.from_netcdf` is an alias for `xarray.open_datatree`, and `az.from_zarr` is a pre-configured partial for Zarr files.

```APIDOC
## `az.from_netcdf` / `az.from_zarr` — Load saved inference results

`az.from_netcdf` is an alias for `xarray.open_datatree`; `az.from_zarr` is a pre-configured partial that sets `engine="zarr"`. Both load existing inference data files saved in NetCDF or Zarr formats, including files originally saved from legacy `arviz.InferenceData` objects.

```python
import arviz as az

# Load a NetCDF file (engine auto-detected or explicit)
dt = az.from_netcdf("inference_results.nc")
dt = az.from_netcdf("inference_results.nc", engine="h5netcdf")

# Load a Zarr store
dt = az.from_zarr("inference_results.zarr")

# Save and reload round-trip example
dt = az.load_arviz_data("centered_eight")
dt.to_netcdf("example.nc", engine="h5netcdf")
dt_reloaded = az.from_netcdf("example.nc")

print(dt_reloaded.groups)
# ('/', '/posterior', '/posterior_predictive', '/log_likelihood',
#  '/sample_stats', '/prior', '/prior_predictive', '/observed_data', '/constant_data')

# Access a specific group as DataTree or Dataset
posterior_dt = dt_reloaded["posterior"]           # DataTree
posterior_ds = dt_reloaded["posterior"].dataset    # Dataset (view)
theta = dt_reloaded["posterior"]["theta"]          # DataArray (unchanged)
```
```

--------------------------------

### Execute Tox Development Tasks

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Run specific tox environments for tasks like testing or code style checks.

```bash
tox -e test-namespace
tox -e check
```

--------------------------------

### Resetting computational backend to 'base'

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Resets the computational backend for ArviZ statistics functions to 'base' by setting `az.rcParams["stats.module"]`.

```python
az.rcParams["stats.module"] = "base"
```

--------------------------------

### Access Sample Statistics

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyMC3_schema_example.ipynb

Retrieves and displays the sampling statistics, such as energy, step size, and divergence information, from the InferenceData object.

```python
idata_pymc3.sample_stats
```

--------------------------------

### Create Feature Branch

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_tutorial.md

Create a new branch for your development work. Always create a feature branch before making changes.

```bash
git checkout -b my-feature
```

--------------------------------

### DataTree Group Navigation and Transformation

Source: https://context7.com/arviz-devs/arviz/llms.txt

Demonstrates how to navigate, filter, and transform ArviZ data structures using xarray's DataTree. Supports accessing groups as DataTrees, Datasets, or DataArrays.

```python
import arviz as az

dt = az.load_arviz_data("centered_eight")

# Check available groups
print(list(dt.children))
# ['posterior', 'posterior_predictive', 'log_likelihood', ...]

# Access a group as DataTree, Dataset, or DataArray
post_dt = dt["posterior"]              # DataTree
post_ds = dt["posterior"].dataset      # Dataset (view)
theta = dt["posterior"]["theta"]       # DataArray

# Apply a function to selected groups (replaces idata.map)
shifted = dt.copy()
shifted.update(
    dt.match("*_predictive").map_over_datasets(lambda ds: ds + 3)
)

# Filter to specific groups
subset = dt.filter(
    lambda node: node.name in ("posterior", "observed_data")
)

# Merge two DataTrees (replaces idata.extend)
dt_new = az.from_dict({"prior": {"alpha": [0.1, 0.2, 0.3]}})
dt.update(dt_new)   # left-merge: dt groups take priority

# Select specific coordinates
dt.sel(school=["Choate", "Deerfield"])
```

--------------------------------

### Set ArviZ Style and Plot with Matplotlib

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Applies a global ArviZ style and then generates a rank plot using the matplotlib backend. Ensure ArviZ data is loaded before plotting.

```python
az.style.use("arviz-vibrant")
dt = az.load_arviz_data("centered_eight")
az.plot_rank(dt, var_names=["mu", "tau"], backend="matplotlib");
```

--------------------------------

### Troubleshoot ArviZ Import Errors

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/getting_started.md

If you encounter import errors, configure logging to INFO level and re-import ArviZ. This can help in diagnosing issues by showing detailed import messages.

```python
import logging
logging.basicConfig(level=logging.INFO)
import arviz as az
```

--------------------------------

### General Intersphinx Reference Syntax in rST

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/syntax_guide.md

Shows the general syntax for creating intersphinx links in reStructuredText, including the optional intersphinx key.

```rst
:type_id:`(intersphinx_key:)object_id`
```

--------------------------------

### Replicate inplace behavior with update

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

To replicate the `inplace=True` behavior of the old `.map` method, create a copy of the DataTree, apply the transformation to the desired groups using `match` and `map_over_datasets`, and then update the original DataTree with the modified groups using `.update`.

```python
shifted_dt = dt.copy()
shifted_dt.update(dt.match("*_predictive").map_over_datasets(lambda ds: ds + 3))
```

--------------------------------

### Generating Random Data for Conversion

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/user_guide/migration_guide.ipynb

Generate random data using numpy's random number generator. This data will be used to demonstrate ArviZ's enhanced converter flexibility.

```python
import numpy as np
rng = np.random.default_rng()
data = rng.normal(size=1000)
```

--------------------------------

### Update Version String for Development

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/how_to_release.md

After a release, update the version string to include the development flag (e.g., `.dev0`) to indicate ongoing development. This follows PEP 440 compliance.

```text
0.13.0.dev0
```

--------------------------------

### General Intersphinx Reference Syntax in MyST Markdown

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/syntax_guide.md

Shows the general syntax for creating intersphinx links in MyST Markdown, including the optional intersphinx key.

```markdown
{type_id}`(intersphinx_key:)object_id`
```

--------------------------------

### Run Tests for Python 3.12 Environment

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/pr_checklist.md

Execute all tests for the Python 3.12 environment using tox. Replace 'py312' with your target Python version if needed (e.g., 'py311', 'py313').

```shell
tox -e py312
```

--------------------------------

### Link to Matplotlib Docs in rST

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/contributing/syntax_guide.md

Demonstrates how to create a link to Matplotlib's Axes.plot documentation using a specific intersphinx key in reStructuredText.

```rst
:meth:`mpl:matplotlib.axes.Axes.plot`
```

--------------------------------

### Access Prior Data in PyStan

Source: https://github.com/arviz-devs/arviz/blob/main/docs/source/schema/PyStan_schema_example.ipynb

Retrieve the prior distributions defined for the PyStan model parameters. This dataset shows the initial beliefs about the parameters before observing data.

```python
idata_stan.prior
```