marimo-notebook

Name: marimo-notebook
Author: pymc-labs

103installs10forks

How to install marimo-notebook?

pymc-labs/python-analytics-skills/marimo-notebook

$npx -y skills add pymc-labs/python-analytics-skills --skill marimo-notebook

Installs into the current project.

Summary

ALWAYS use when: creating/editing marimo notebooks, working with any .py file containing @app.cell decorators, building reactive Python notebooks, doing exploratory data analysis in notebook form, converting Jupyter (.ipynb) to marimo, or when user mentions "marimo", "reactive no

Files · 1

View on GitHub

SKILL.md

1	# Marimo Notebooks
2
3	Marimo notebooks are reactive Python notebooks stored as pure `.py` files. Cells auto-execute when dependencies change, modeled as a directed acyclic graph (DAG).
4
5	## Core Concepts
6
7	### Reactivity Model
8	- marimo uses static analysis to build a dependency graph from variable references and definitions
9	- When a cell runs, all cells referencing its defined variables automatically re-run
10	- Execution order follows the dependency graph, not visual cell order
11	- Each global variable must be defined by exactly one cell
12	- marimo does not track object mutations (like `list.append()`)—mutate in the same cell that creates the object, or create new variables
13
14	### Avoiding Variable Name Conflicts
15	Each global variable must be defined by exactly one cell. Two strategies:
16
17	1. Wrap code in functions (preferred for reusable patterns):
18	```python
19	@app.cell
20	def _(data):
21	def compute_mean_with_new_col(df):
22	temp = df.copy()
23	temp["new_col"] = temp["x"] * 2
24	return temp.mean()
25
26	return (compute_mean_with_new_col(data),)
27	```
28
29	2. Use meaningful, unique variable names:
30	```python
31	@app.cell
32	def _(model1_data):
33	model1_transformed = model1_data.copy()
34	model1_transformed["new_col"] = model1_transformed["x"] * 2
35	return (model1_transformed,)
36	```
37
38	Never use underscore prefixes to generate unique variable names. No exceptions.
39
40	## Notebook Structure
41
42	```python
43	import marimo
44
45	__generated_with = "0.10.0"
46	app = marimo.App(width="medium")
47
48	@app.cell
49	def _():
50	import marimo as mo
51	return (mo,)
52
53	@app.cell
54	def _(mo):
55	mo.md("# Hello")
56	return
57
58	if __name__ == "__main__":
59	app.run()
60	```
61
62	Key rules:
63	- Each cell is a function decorated with `@app.cell`
64	- Variables shared by returning tuples: `return (var1, var2,)`
65	- Cells receive variables as parameters: `def _(mo, df):`
66	- Execution order follows dependency graph, not position
67	- Name cells descriptively for CellTour targeting: `def model_specification():`
68
69	## CLI Commands
70
71	```bash
72	# Create & Edit
73	marimo new # Create new notebook
74	marimo edit notebook.py # Open editor
75	marimo edit notebook.py --watch # Live reload on file changes
76
77	# Run as App
78	marimo run notebook.py # Run as app (code hidden by default)
79	marimo run notebook.py --include-code # Show code in app view
80
81	# Convert
82	marimo convert notebook.ipynb -o notebook.py # Jupyter to marimo
83
84	# Export
85	marimo export html notebook.py -o out.html # Static HTML
86	marimo export ipynb notebook.py -o out.ipynb # To Jupyter
87
88	# Validate
89	marimo check notebook.py # Lint and validate
90	marimo check notebook.py --fix # Auto-fix issues
91	```
92
93	## Code Visibility in Run Mode
94
95	CRITICAL FOR TUTORIALS: By default, `marimo run` hides code. Use `mo.show_code()` to display it.
96
97	### mo.show_code() - Per-Cell Display
98
99	IMPORTANT: Call `mo.show_code()` as a statement on its own line, NOT in the return statement.
100
101	```python
102	@app.cell
103	def model_definition(mo, pm, X, y):
104	with pm.Model() as model:
105	alpha = pm.Normal("alpha", mu=0, sigma=10)
106	beta = pm.Normal("beta", mu=0, sigma=10)
107	mu = alpha + beta * X
108	pm.Normal("y", mu=mu, sigma=1, observed=y)
109
110	# Show this cell's code alongside its output
111	mo.show_code(model, position="above")
112	return (model,)
113	```
114
115	WRONG vs RIGHT patterns:
116	```python
117	# WRONG - do not put mo.show_code() in return statement
118	return mo.show_code(result)
119
120	# RIGHT - call as statement, then return separately
121	mo.show_code(result, position="above")
122	return (result,)
123	```
124
125	- `position="above"` shows code first, then output (best for tutorials)
126	- `position="below"` shows output first, then code (default)
127
128	## Markdown with mo.md()
129
130	```python
131	@app.cell
132	def _(mo):
133	mo.md(r"""
134	# Title
135
136	Interpolate Python: {slider}
137
138	LaTeX: $f(x) = e^x$
139
140	$$\int_0^\infty e^{-x^2} dx = \frac{\sqrt{\pi}}{2}$$
141
142	Icons: ::lucide:rocket:: or ::mdi:home::
143	""")
144	return
145	```
146
147	- Use raw strings (`r"""..."""`) for LaTeX
148	- Interpolate UI elements: `f"Value: {slider}"`
149	- For complex objects: `f"Plot: {mo.as_html(fig)}"`
150
151	## UI Components (mo.ui.*)
152
153	### Basic Inputs
154	```python
155	slider = mo.ui.slider(0, 100, value=50, label="Value")
156	number = mo.ui.number(0, 100, value=50)
157	text = mo.ui.text(value="", placeholder="Enter text")
158	checkbox = mo.ui.checkbox(value=F

People who liked this also viewed

6 picks

Type

marimo-notebook

103installs10forks

View on GitHub

How to install marimo-notebook?

pymc-labs/python-analytics-skills/marimo-notebook

$npx -y skills add pymc-labs/python-analytics-skills --skill marimo-notebook

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Marimo Notebooks
2
3	Marimo notebooks are reactive Python notebooks stored as pure `.py` files. Cells auto-execute when dependencies change, modeled as a directed acyclic graph (DAG).
4
5	## Core Concepts
6
7	### Reactivity Model
8	- marimo uses static analysis to build a dependency graph from variable references and definitions
9	- When a cell runs, all cells referencing its defined variables automatically re-run
10	- Execution order follows the dependency graph, not visual cell order
11	- Each global variable must be defined by exactly one cell
12	- marimo does not track object mutations (like `list.append()`)—mutate in the same cell that creates the object, or create new variables
13
14	### Avoiding Variable Name Conflicts
15	Each global variable must be defined by exactly one cell. Two strategies:
16
17	1. Wrap code in functions (preferred for reusable patterns):
18	```python
19	@app.cell
20	def _(data):
21	def compute_mean_with_new_col(df):
22	temp = df.copy()
23	temp["new_col"] = temp["x"] * 2
24	return temp.mean()
25
26	return (compute_mean_with_new_col(data),)
27	```
28
29	2. Use meaningful, unique variable names:
30	```python
31	@app.cell
32	def _(model1_data):
33	model1_transformed = model1_data.copy()
34	model1_transformed["new_col"] = model1_transformed["x"] * 2
35	return (model1_transformed,)
36	```
37
38	Never use underscore prefixes to generate unique variable names. No exceptions.
39
40	## Notebook Structure
41
42	```python
43	import marimo
44
45	__generated_with = "0.10.0"
46	app = marimo.App(width="medium")
47
48	@app.cell
49	def _():
50	import marimo as mo
51	return (mo,)
52
53	@app.cell
54	def _(mo):
55	mo.md("# Hello")
56	return
57
58	if __name__ == "__main__":
59	app.run()
60	```
61
62	Key rules:
63	- Each cell is a function decorated with `@app.cell`
64	- Variables shared by returning tuples: `return (var1, var2,)`
65	- Cells receive variables as parameters: `def _(mo, df):`
66	- Execution order follows dependency graph, not position
67	- Name cells descriptively for CellTour targeting: `def model_specification():`
68
69	## CLI Commands
70
71	```bash
72	# Create & Edit
73	marimo new # Create new notebook
74	marimo edit notebook.py # Open editor
75	marimo edit notebook.py --watch # Live reload on file changes
76
77	# Run as App
78	marimo run notebook.py # Run as app (code hidden by default)
79	marimo run notebook.py --include-code # Show code in app view
80
81	# Convert
82	marimo convert notebook.ipynb -o notebook.py # Jupyter to marimo
83
84	# Export
85	marimo export html notebook.py -o out.html # Static HTML
86	marimo export ipynb notebook.py -o out.ipynb # To Jupyter
87
88	# Validate
89	marimo check notebook.py # Lint and validate
90	marimo check notebook.py --fix # Auto-fix issues
91	```
92
93	## Code Visibility in Run Mode
94
95	CRITICAL FOR TUTORIALS: By default, `marimo run` hides code. Use `mo.show_code()` to display it.
96
97	### mo.show_code() - Per-Cell Display
98
99	IMPORTANT: Call `mo.show_code()` as a statement on its own line, NOT in the return statement.
100
101	```python
102	@app.cell
103	def model_definition(mo, pm, X, y):
104	with pm.Model() as model:
105	alpha = pm.Normal("alpha", mu=0, sigma=10)
106	beta = pm.Normal("beta", mu=0, sigma=10)
107	mu = alpha + beta * X
108	pm.Normal("y", mu=mu, sigma=1, observed=y)
109
110	# Show this cell's code alongside its output
111	mo.show_code(model, position="above")
112	return (model,)
113	```
114
115	WRONG vs RIGHT patterns:
116	```python
117	# WRONG - do not put mo.show_code() in return statement
118	return mo.show_code(result)
119
120	# RIGHT - call as statement, then return separately
121	mo.show_code(result, position="above")
122	return (result,)
123	```
124
125	- `position="above"` shows code first, then output (best for tutorials)
126	- `position="below"` shows output first, then code (default)
127
128	## Markdown with mo.md()
129
130	```python
131	@app.cell
132	def _(mo):
133	mo.md(r"""
134	# Title
135
136	Interpolate Python: {slider}
137
138	LaTeX: $f(x) = e^x$
139
140	$$\int_0^\infty e^{-x^2} dx = \frac{\sqrt{\pi}}{2}$$
141
142	Icons: ::lucide:rocket:: or ::mdi:home::
143	""")
144	return
145	```
146
147	- Use raw strings (`r"""..."""`) for LaTeX
148	- Interpolate UI elements: `f"Value: {slider}"`
149	- For complex objects: `f"Plot: {mo.as_html(fig)}"`
150
151	## UI Components (mo.ui.*)
152
153	### Basic Inputs
154	```python
155	slider = mo.ui.slider(0, 100, value=50, label="Value")
156	number = mo.ui.number(0, 100, value=50)
157	text = mo.ui.text(value="", placeholder="Enter text")
158	checkbox = mo.ui.checkbox(value=F

People who liked this also viewed

6 picks

Type