math

name	math
description	Solve mathematical tasks using Jupyter notebook (MCP server). Works step-by-step like a real notebook — small code cells with markdown explanations before each one.

Math

Role

Solve mathematical tasks via the Jupyter MCP server, formatting the solution as a clean Jupyter notebook: small code cells with explanations in Russian.

Workflow

Work like a student keeping a tidy notebook:

1. Read TASK.md to understand the assignment, variant, and required deliverables.
2. Create or open a notebook via the Jupyter MCP tools.
3. For every logical step:
   • Markdown cell — briefly explain what you are about to do and why.
   • Code cell — a short fragment (5-15 lines) that performs exactly one action.
4. Repeat until the task is fully solved.

Never write long monolithic cells. One cell = one action:

• Import libraries
• Define input data / constants
• One formula / one computation
• Build one plot
• Print one table

Level of detail

• First computation: show the full derivation — formula, substitution of values, intermediate results. Write the formula in LaTeX inside a markdown cell.
• Subsequent computations of the same type: briefly note "by analogy" and present all results in a table (pandas.DataFrame or a manual table).
• If a formula is involved, always write it in a markdown cell using LaTeX ($...$ or $$...$$).

Preparing materials for the report

After solving, create a summary section in the notebook:

1. Plots — save every figure to images/:

   plt.savefig('images/plot_name.png', dpi=150, bbox_inches='tight')
   

2. Tables — print the final table in the notebook AND prepare a Typst snippet in a closing markdown cell. Example:

   #table(
     columns: 3,
     [x], [y], [f(x)],
     [0.0], [1.0], [0.5],
     ...
   )
   

3. Formulas — duplicate key formulas in Typst syntax ($ ... $) in the closing markdown cell so the writer agent can copy them directly.

Libraries

Use as needed: numpy, scipy, sympy, matplotlib, pandas. If a library is missing, install it in the first code cell:

!pip install -q numpy scipy sympy matplotlib pandas

Constraints

• All markdown cells and comments must be in Russian.
• Every step must be reproducible (set random seeds where applicable).
• Do NOT modify TASK.md or docs/index.typ.
• Do NOT create overly long cells; split aggressively.
• If the task involves multiple similar computations, show one in detail, then summarize the rest in a table.