Write a Python script that charts your data

# Role
You write short Python scripts with pandas and matplotlib that turn a spreadsheet or CSV file into one clear, accurate chart.

# Goal
Write a script that reads my file and draws the chart described below, for {{audience}}, saved as an image I can put in {{destination}}, so that the chart can be redrawn the same way each time the data changes.

# Success criteria
- It reads the columns by the exact names given below and never types in data values.
- A reader can understand the chart without me: a title, axis labels with units, font sizes readable at the final size, and a legend only when there is more than one series.
- Bar and column charts start at zero, and bars are sorted by value unless the categories have a natural order such as months.
- Dates are parsed as dates, and numbers show thousands separators.
- It runs with python chart.py after I change only the paths at the top.

# Constraints
- Use only pandas, matplotlib and the standard library, plus openpyxl if the file is Excel.
- Do not drop, fill or change rows silently. If rows are excluded (blanks, dates that cannot be read), print how many and why.
- Use one color for a single series, and a colorblind-safe palette (for example Okabe-Ito) for several. Avoid 3D effects, shadows and pie charts with more than five slices.
- Avoid clever one-liners; the next person to change it may not be a programmer.

# Output
1. The script in one code block, with INPUT_PATH, OUTPUT_PATH, the column names and the chart size as constants at the top, and a comment above each step.
2. It saves to {{output_format}} with dpi=200 and bbox_inches="tight", and prints the summarized table it plotted, so I can compare a few values with my spreadsheet.
3. Below the script, up to six lines on how to install the packages and run it.

For example, a title that works reads: "Monthly orders by region, January to June". That example shows the form only; write the title from my description.

# Stop rules
If the sample rows do not match the column names I gave, or a column mixes formats (dates written two ways, numbers stored as text with currency signs), say so before the script and handle it explicitly in the code. If the chart I asked for would mislead or be hard to read, such as a line chart across unordered categories, say so in one line and suggest a better type, then write the one I asked for.

# Context
File type and path: {{data_file}}
Columns and what they hold: {{columns}}
The chart (type, what goes on each axis, grouping, and whether values are summed or averaged): {{chart_wanted}}
Python version and what I can install: {{setup}}

Sample rows, including the header:
"""
{{sample_rows}}
"""