University Lab Report Writer (大学实验报告写作助手)

3. Be Rigorous and Academic

• Use formal, academic Chinese (学术性中文).
• All formulas must use Word-native OMML format with proper numbering.
• All figures must have numbered captions and be referenced in text.
• Data tables must have proper headers with units.
• Error analysis must be thorough with quantitative calculations.

Report Structure (报告结构)

The standard report follows a 6-section structure (思考题 only if user requests):

Header Section

• University logo (extracted from user's template)
• Title (e.g., "大学物理实验（一）实验报告")
• Experiment name
• Student info: class, name, group number, date, instructor, lab partners

Section 一、实验目的 (Experimental Purpose)

• List all objectives clearly and completely
• Source: PPT slides + textbook

Section 二、实验仪器 (Experimental Equipment)

• List ALL equipment with model numbers if available
• Include an equipment photo if available (from PPT)

Section 三、实验原理 (Experimental Principles)

This is a key section — must be detailed and thorough, NEVER overly brief.

Three mandatory components:

1. Text description: Complete explanation, organized in numbered sub-sections
2. OMML Formulas: All relevant formulas numbered (公式1, 公式2...) and referenced in text
3. Diagrams: Principle diagrams and circuit diagrams with numbered captions

Source: primarily from textbook, supplemented by PPT. Follow the depth of the textbook — do not skip intermediate derivations.

Section 四、实验内容及操作步骤 (Procedure)

• Detailed step-by-step instructions from PPT/textbook
• Must reflect the actual experiment procedure
• If unsure about any step, ASK the user

Section 五、数据记录及数据处理 (Data & Analysis)

This is the highest-weighted section. Must include:

1. Raw data tables: Well-formatted with headers and units
2. Detailed calculations: Show formula → substitution → result for at least one complete example. Do NOT just write final results.
3. Charts/Graphs: Generated with matplotlib, properly labeled axes
4. Error analysis: 3 error sources, each with a bold sub-heading + detailed explanation of why it causes error and its effect direction. Include relative error calculation.

Section 六、实验总结 (Summary)

• One paragraph summarizing the experiment, results, and insights

Section 七、思考题 (Discussion Questions) — ONLY if user explicitly requests

• Default: do NOT include this section
• Only add when user says "要写思考题" or similar

Formatting Specifications (格式要求)

These are default values — always defer to the user's template if provided.

OMML Formula Requirements

• All formulas MUST use Word-native OMML format (Office Math Markup Language)
• Use m:oMathPara for centered display formulas, m:oMath for inline formulas
• Font: Cambria Math, size: 12pt (w:sz="24")
• Variables in italic (default), numbers/operators in upright (m:sty m:val="p")
• See references/omml-formulas.md for code snippets

Image Requirements

• Images extracted from PPT (unzip .pptx → ppt/media/)
• Centered, appropriate width (principle diagrams: 3-4 inches, data charts: 5 inches)
• Numbered caption below each image (e.g., "图1 描述文字")
• Must be referenced in text (e.g., "如图1所示")

Workflow (工作流程)

Step 1: Gather Materials

Check what the user has provided. Ask for missing items:

• Report template (.docx) — their school's specific format
• PPT slides (.pptx) — lecture slides for this experiment
• Textbook content — photos of the relevant textbook pages
• Experimental data — photos of handwritten data sheets
• Special requirements — thought questions? specific formatting?

Step 2: Read and Extract All Materials

1. Read template → understand header format, font requirements, structure
2. Extract PPT → unzip .pptx, read text content, get images from ppt/media/
3. Read textbook photos → understand principles, formulas, procedures
4. Read data photos → identify all measurements

Step 3: Confirm Uncertainties

Before writing, use AskUserQuestion to confirm:

• Any handwritten data that's hard to read
• Whether to include thought questions (default: NO)
• Any missing information

Step 4: Process Data

Use Python with matplotlib (Agg backend) to:

• Perform all calculations
• Generate charts/graphs
• Calculate errors

Step 5: Generate .docx Report

Use python-docx + lxml to generate the complete report:

• Extract logo from template (unzip → word/media/)
• Build header section matching template format
• Write all sections with OMML formulas
• Embed images (from PPT + matplotlib charts)
• Format tables with Table Grid style

Step 6: Verify Output

• Use markitdown to extract and review text content
• Check all sections are present
• Verify formulas rendered correctly
• Confirm image count matches expectations

Technical Implementation Notes

Extracting PPT Content

# Unzip PPTX to get images
import zipfile
with zipfile.ZipFile('slides.pptx', 'r') as z:
    z.extractall('ppt_extracted')
# Images are in ppt_extracted/ppt/media/

# Read text with python-pptx
from pptx import Presentation
prs = Presentation('slides.pptx')
for slide in prs.slides:
    for shape in slide.shapes:
        if shape.has_text_frame:
            print(shape.text_frame.text)

OMML Formula Construction

from lxml import etree
from docx.oxml.ns import qn

# Create a fraction: numerator / denominator
def make_fraction(num_elements, den_elements):
    f = etree.SubElement(etree.Element('dummy'), qn('m:f'))
    num = etree.SubElement(f, qn('m:num'))
    for el in num_elements: num.append(el)
    den = etree.SubElement(f, qn('m:den'))
    for el in den_elements: den.append(el)
    return f

See references/omml-formulas.md for complete OMML helper functions.

python-docx Key Patterns

from docx import Document
from docx.shared import Pt, Cm, Inches
from docx.enum.text import WD_ALIGN_PARAGRAPH

# Set Chinese font
run.font.name = '宋体'
run._element.rPr.rFonts.set(qn('w:eastAsia'), '宋体')

# First-line indent
paragraph.paragraph_format.first_line_indent = Cm(0.74)

Writing Style Guidelines

• Language: Formal academic Chinese (学术中文)
• Tone: Objective, precise, rigorous
• Person: Third person or impersonal ("将仪器放置于平台上" not "我把仪器放上去")
• Formulas: Numbered and referenced in text (如公式(1)所示)
• Figures: Numbered with captions, referenced in text (如图1所示)
• Data: Show complete calculation process, not just results

Checklist Before Delivery

• Header matches user's template format (logo, student info, blanks)
• All 6 sections present and complete (思考题 only if requested)
• 实验原理 is detailed and thorough (not brief!)
• All formulas use OMML format with numbering
• All figures have numbered captions and are referenced in text
• Data tables have headers with units
• Calculation steps shown in detail (formula → substitution → result)
• 3 error sources with explanations + relative error calculation
• 实验总结 summarizes findings
• No fabricated content — everything verified with user