pythesis-plot（假设图）

PyThesisPlot

Python scientific plotting workflow tool supporting the complete process from data upload to figure generation for academic publications.

Workflow

[User Uploads Data] → [Auto-save to output dir] → [Data Analysis]
                                           ↓
[Generate Images to output dir] ← [Code Generation] ← [User Confirms Scheme]

Required Steps

1. Data Reception: User uploads data file (txt/md/xlsx/csv)
2. Auto-save: Rename to timestamp-original_filename, save to output/YYYYMMDD-filename/
3. Data Analysis: Analyze dimensions, types, statistical features, column relationships
4. Chart Recommendations: Recommend chart schemes based on data characteristics (type, quantity, layout)
5. User Confirmation: Display analysis report, must wait for user confirmation before generation
6. Generation & Delivery: Python code + chart images, save to same output directory

Core Scripts

1. Main Workflow Script

python scripts/workflow.py --input data.csv --output-dir output/

2. Data Analysis

python scripts/data_analyzer.py --input data.csv

Output: Data characteristics report + chart recommendation scheme

3. Chart Generation

python scripts/plot_generator.py --config plot_config.json --output-dir output/

File Management Standards

Directory Structure

output/
└── 20250312-145230-data.csv/          # Named with timestamp + filename
    ├── 20250312-145230-data.csv       # Original data file (renamed)
    ├── analysis_report.md             # Data analysis report
    ├── plot_config.json               # Chart configuration (generated after user confirmation)
    ├── 20250312-145230_plot.py        # Generated Python code
    ├── 20250312-145230_fig1_line.png  # Chart (PNG image)
    └── 20250312-145230_fig2_bar.png

Naming Conventions

Usage

Scenario 1: Complete Workflow

When user uploads a data file:

1. Auto-save File

   # Rename and save to output/{timestamp}-{filename}/
   save_uploaded_file(input_file, output_base="output/")

1. Execute Data Analysis

   # Analyze data characteristics, generate report
   python scripts/data_analyzer.py --input output/20250312-data/data.csv

1. Display Analysis Report to User

   ## Data Analysis Report
   
   ### Data Overview
   - File: data.csv
   - Dimensions: 120 rows × 5 columns
   - Types: 3 numeric + 2 categorical columns
   
   ### Column Details
   | Column | Type | Description |
   |-----|------|-----|
   | date | datetime | 2023-01 to 2023-12 |
   | sales | numeric | mean=1250, std=320 |
   | region | categorical | 4 categories: N/S/E/W |
   
   ### Chart Recommendations
   Based on data characteristics, the following schemes are recommended:
   
   **Scheme 1: Time Trend Analysis** ⭐Recommended
   - Chart Type: Line plot
   - Content: Sales trend over time
   - Reason: Time series data, most intuitive for showing trends
   
   **Scheme 2: Regional Comparison**
   - Chart Type: Grouped bar chart
   - Content: Sales comparison across regions
   - Reason: Categorical comparison, suitable for showing differences
   
   **Scheme 3: Comprehensive Dashboard**
   - Chart Type: 2×2 subplot layout
   - Includes: Trend line + Bar chart + Box plot + Correlation heatmap
   - Reason: Rich data dimensions, comprehensive display
   
   Please tell me what you want:
   - "Generate schemes 1 and 2"
   - "Generate all"
   - "Modify scheme 3..." (provide your modification suggestions)

1. Wait for User Confirmation ⚠️ Critical Step

- User may say: "Generate scheme 1" / "Generate all" / "Modify XX..." - Must wait for explicit instruction before entering generation phase

1. Generate and Save

   # Generate Python code
   python scripts/plot_generator.py --config plot_config.json
   
   # Output to same directory
   output/20250312-data/
   ├── 20250312-145230_plot.py        # Code
   ├── 20250312-145230_fig1_line.png  # Chart
   └── 20250312-145230_fig2_bar.png

Scenario 2: Data Analysis Only

python scripts/data_analyzer.py --input data.csv --output report.md

Scenario 3: Generate from Config

python scripts/plot_generator.py --config config.json --output-dir ./

Chart Recommendation Logic

Supported File Formats

• CSV: .csv (Recommended)
• Excel: .xlsx, .xls
• Text: .txt, .md (table format)

Dependencies

pandas >= 1.3.0
matplotlib >= 3.5.0
seaborn >= 0.11.0
openpyxl >= 3.0.0  # Excel support
numpy >= 1.20.0
scipy >= 1.7.0

Reference Documents

• Workflow Guide - Complete workflow instructions
• Chart Types - Detailed chart type descriptions
• Style Guide - Color schemes, fonts, size standards
• Examples - Complete code examples

Important Notes

1. User confirmation is mandatory: Must wait for user confirmation after analysis, cannot generate directly
2. Unified file management: All output files saved to same output/{timestamp}-{filename}/ directory
3. High-resolution output: Generate PNG at 300 DPI (suitable for publication)
4. Code traceability: Generated Python code also saved to same directory for user modification
5. Academic style: Charts follow top journal standards (Nature/Science/Lancet style)