supervised-machine-learning/python_in_posit_cloud.qmd at main · laser-institute/supervised-machine-learning · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
---
title: "Using Python in Posit Cloud"
format:
  html:
    toc: true
    number-sections: true
---

# Short Answer

In each Posit Cloud project, create and activate a project-scoped Python virtual environment, install your needed Python packages there, tell Posit (RStudio) which Python to use (via Project Options → Python, `RETICULATE_PYTHON`, or `reticulate::use_virtualenv()`), and then work in Quarto (.qmd) documents (Knitr + reticulate for mixed R/Python or Jupyter engine for Python-first) or full Jupyter notebooks. Capture dependencies in `requirements.txt` (or `environment.yml`), and publish/share with `quarto publish posit-cloud` (or `rsconnect-python`).

# 1. Project Setup (once per project)

1. **Create a project-local virtual environment**:

   ```bash
   python3 -m venv env
   ```

2. **Activate the environment & install packages**:

   ```bash
   source env/bin/activate
   pip install jupyter pandas matplotlib # etc.
   ```

3. **Bind Posit/RStudio to the environment**:

   - **GUI**: Tools → Project Options → Python → select `env/bin/python`
   - **Env var**: add `RETICULATE_PYTHON=env/bin/python` to `.Renviron`
   - **R code**:
     ```r
     reticulate::use_virtualenv("env", required = TRUE)
     ```

4. **(Optional) Install from R**:

   ```r
   reticulate::py_install("dplyr")
   # or
   reticulate::virtualenv_install("env", "scikit-learn")
   ```

5. **Verify interpreter**:

   ```r
   reticulate::py_config()
   ```

# 2. Daily Workflows

## A. Quarto documents with Python chunks

Use executable chunks in your .qmd:

```{python}
#| echo: true
import pandas as pd
import numpy as np
```

Add `jupyter: python3` (or your kernel) in the YAML to run via Jupyter.

## B. Mixed R + Python with reticulate (Knitr engine)

```{python}
# Python chunk
array = np.arange(10)
```

Exchange objects between R and Python seamlessly:

```{r}
py$array * 2
```

## C. Python in the Console

- Send lines from a `.py` file (Cmd/Ctrl+Enter)
- Drop into Python REPL
- View matplotlib plots in the Plots pane

## D. Full Jupyter Notebooks

- Create a new project with the Jupyter engine
- Work directly in `.ipynb` notebooks

# 3. Reproducibility & Dependencies

- **Capture packages**:
  ```bash
  pip freeze > requirements.txt
  ```
- **Conda alternative**:
  ```bash
  conda env export > environment.yml
  ```
- **R + Python together** with renv:
  ```r
  renv::init()
  renv::use_python()
  renv::snapshot()
  ```

# 4. Sharing & Publishing

- **Publish**:
  ```bash
  quarto publish posit-cloud
  ```
- **Connect Cloud**:
  - Include `.ipynb` and `requirements.txt`

# 5. Common Gotchas

| Symptom                          | Cause                                     | Fix                                                                   |
| -------------------------------- | ----------------------------------------- | --------------------------------------------------------------------- |
| Quarto can’t find Python         | Env not activated; wrong Python in Quarto | Activate venv; set `QUARTO_PYTHON`; run `quarto check jupyter`        |
| reticulate using system Python   | `RETICULATE_PYTHON` not set               | Set in Project Options or `.Renviron`; restart; confirm `py_config()` |
| Missing packages when publishing | `requirements.txt` not updated            | `pip freeze > requirements.txt` before publish                        |

# 6. Next Steps

- Are you mixing small Python snippets into R teaching materials, building Python-first notebooks, or setting up a reproducible workshop environment? Let me know which workflow fits your needs best!