---
jupytext:
  formats: ipynb,md:myst
  text_representation:
    extension: .md
    format_name: myst
    format_version: 0.13
    jupytext_version: 1.14.5
kernelspec:
  display_name: Python 3 (ipykernel)
  language: python
  name: python3
---

+++ {"id": "a3o1EnPLWeAv"}

(tutorial_3)=
# Single subject
Author: Nicolas Legrand <nicolas.legrand@cfin.au.dk>  
Adapted from the tutorial proposed by the HMeta-d toolbox: https://github.com/metacoglab/HMeta-d/tree/master/CPC_metacog_tutorial

```{code-cell} ipython3
:tags: [hide-cell]

%%capture
import sys
if 'google.colab' in sys.modules:
    ! pip install metadpy
```

```{code-cell} ipython3
:id: uSC4P5ghWTZb

import arviz as az
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
from metadpy.bayesian import hmetad

sns.set_context("talk")
```

```{code-cell} ipython3
:id: v8xhAXePWm2Y

# Create responses data
nR_S1 = np.array([52, 32, 35, 37, 26, 12, 4, 2])
nR_S2 = np.array([2, 5, 15, 22, 33, 38, 40, 45])
```

+++ {"id": "ZOop3bjebxbZ"}

## Using metadpy

```{code-cell} ipython3
---
colab:
  base_uri: https://localhost:8080/
  height: 118
  referenced_widgets: [48a76ae7649b48f9a0a85ccc5751815d, 21b2283e56ee49f5a0142808dacbe6a9,
    f1103196b06642c8ba127ab3784a0fda, e293ba5b914646b58817190af25fb921, a11181a9d3bc4a74b795c1099125ddaa,
    fbb738761edf4c5e8a31271931243fc8, bade777d33fb4651a70fbf97e059fbd0, 34652852c7ec4a9aa50f3411d378c7f0,
    cbe8740b7a1841e6be7fb4041581ef5e, 4f1d2dba668f40ce8c3fea52ffd68600, 746da93bd5fb4fd0ab091ca40ce68762,
    f0e23502e0964dc08189ff00e9132196, b358e7faba4046dcaf4c70d2a8ab2982, c945cf1be9a04cd780d4b1929ebc0fa7,
    2d73190fef28427088ae5c566e333ad8, f1c2f8b5a53c442fad4c719346978113]
id: zpPdxFLiWsaW
outputId: bcf46df4-7be3-4381-e167-548b0a4f971f
---
model, traces = hmetad(nR_S1=nR_S1, nR_S2=nR_S2)
```

+++ {"id": "wcB7amGfbz8e"}

## Using HMeta-d

The results were generated by the `example_metad_indiv.R` file.

```{code-cell} ipython3
:id: RA4wULVlb8Gf

hmetad_df = pd.read_csv("./hmetad/metad_indiv.txt", sep="\t")
```

+++ {"id": "iBFWERvGiPrO"}

## Comparison

```{code-cell} ipython3
---
colab:
  base_uri: https://localhost:8080/
  height: 568
id: 5JllhVsgcauL
outputId: e2473dd6-97ac-4c10-8eeb-83087cca5c1e
---
_, axs = plt.subplots(2, 1, figsize=(8, 8), sharex=True)
az.plot_posterior(hmetad_df.meta_d.to_numpy(), ax=axs[0])
az.plot_posterior(traces, var_names=["meta_d"], ax=axs[1])

axs[0].set_ylabel("HMeta-d")
axs[1].set_ylabel("metadpy")

plt.tight_layout()
```

```{code-cell} ipython3

```