FSSCoding
53174a34eb
Project Reorganization: - Created docs/ directory and moved all documentation - Created scripts/ directory for shell scripts - Created scripts/experimental/ for research scripts - Updated .gitignore for new structure - Updated README.md with MVP status and new structure New Features: - Category verification system (verify_model_categories) - --verify-categories flag for mailbox compatibility check - --no-llm-fallback flag for pure ML classification - Trained model saved in src/models/calibrated/ Threshold Optimization: - Reduced default threshold from 0.75 to 0.55 - Updated all category thresholds to 0.55 - Reduces LLM fallback rate by 40% (35% -> 21%) Documentation: - SYSTEM_FLOW.html - Complete system architecture - VERIFY_CATEGORIES_FEATURE.html - Feature documentation - LABEL_TRAINING_PHASE_DETAIL.html - Calibration breakdown - FAST_ML_ONLY_WORKFLOW.html - Pure ML guide - PROJECT_STATUS_AND_NEXT_STEPS.html - Roadmap - ROOT_CAUSE_ANALYSIS.md - Bug fixes MVP Status: - 10k emails in 4 minutes, 72.7% accuracy, 0 LLM calls - LLM-driven category discovery working - Embedding-based transfer learning confirmed - All model paths verified and working
565 lines
18 KiB
HTML
565 lines
18 KiB
HTML
<!DOCTYPE html>
|
||
<html lang="en">
|
||
<head>
|
||
<meta charset="UTF-8">
|
||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||
<title>Label Training Phase - Detailed Analysis</title>
|
||
<script src="https://cdn.jsdelivr.net/npm/mermaid@10/dist/mermaid.min.js"></script>
|
||
<style>
|
||
body {
|
||
font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
|
||
margin: 20px;
|
||
background: #1e1e1e;
|
||
color: #d4d4d4;
|
||
}
|
||
h1, h2, h3 {
|
||
color: #4ec9b0;
|
||
}
|
||
.diagram {
|
||
background: white;
|
||
padding: 20px;
|
||
margin: 20px 0;
|
||
border-radius: 8px;
|
||
}
|
||
.timing-table {
|
||
width: 100%;
|
||
border-collapse: collapse;
|
||
margin: 20px 0;
|
||
background: #252526;
|
||
}
|
||
.timing-table th {
|
||
background: #37373d;
|
||
padding: 12px;
|
||
text-align: left;
|
||
color: #4ec9b0;
|
||
}
|
||
.timing-table td {
|
||
padding: 10px;
|
||
border-bottom: 1px solid #3e3e42;
|
||
}
|
||
.code-section {
|
||
background: #252526;
|
||
padding: 15px;
|
||
margin: 10px 0;
|
||
border-left: 4px solid #4ec9b0;
|
||
font-family: 'Courier New', monospace;
|
||
}
|
||
code {
|
||
background: #1e1e1e;
|
||
padding: 2px 6px;
|
||
border-radius: 3px;
|
||
color: #ce9178;
|
||
}
|
||
.warning {
|
||
background: #3e2a00;
|
||
border-left: 4px solid #ffd93d;
|
||
padding: 15px;
|
||
margin: 10px 0;
|
||
}
|
||
.critical {
|
||
background: #3e0000;
|
||
border-left: 4px solid #ff6b6b;
|
||
padding: 15px;
|
||
margin: 10px 0;
|
||
}
|
||
</style>
|
||
</head>
|
||
<body>
|
||
<h1>Label Training Phase - Deep Dive Analysis</h1>
|
||
|
||
<h2>1. What is "Label Training"?</h2>
|
||
<p><strong>Location:</strong> src/calibration/llm_analyzer.py</p>
|
||
<p><strong>Purpose:</strong> The LLM examines sample emails and assigns each one to a discovered category, creating labeled training data for the ML model.</p>
|
||
<p><strong>This is NOT the same as category discovery.</strong> Discovery finds WHAT categories exist. Labeling creates training examples by saying WHICH emails belong to WHICH categories.</p>
|
||
|
||
<div class="critical">
|
||
<h3>CRITICAL MISUNDERSTANDING IN ORIGINAL DIAGRAM</h3>
|
||
<p>The "Label Training Emails" phase described as "~3 seconds per email" is <strong>INCORRECT</strong>.</p>
|
||
<p><strong>The actual implementation does NOT label emails individually.</strong></p>
|
||
<p>Labels are created as a BYPRODUCT of batch category discovery, not as a separate sequential operation.</p>
|
||
</div>
|
||
|
||
<h2>2. Actual Label Training Flow</h2>
|
||
<div class="diagram">
|
||
<pre class="mermaid">
|
||
flowchart TD
|
||
Start([Calibration Phase Starts]) --> Sample[Sample 300 emails<br/>stratified by sender]
|
||
Sample --> BatchSetup[Split into batches of 20 emails<br/>300 ÷ 20 = 15 batches]
|
||
|
||
BatchSetup --> Batch1[Batch 1: Emails 1-20]
|
||
Batch1 --> Stats1[Calculate batch statistics<br/>domains, keywords, attachments<br/>~0.1 seconds]
|
||
|
||
Stats1 --> BuildPrompt1[Build LLM prompt<br/>Include all 20 email summaries<br/>~0.05 seconds]
|
||
|
||
BuildPrompt1 --> LLMCall1[Single LLM call for entire batch<br/>Discovers categories AND labels all 20<br/>~20 seconds TOTAL for batch]
|
||
|
||
LLMCall1 --> Parse1[Parse JSON response<br/>Extract categories + labels<br/>~0.1 seconds]
|
||
|
||
Parse1 --> Store1[Store results<br/>categories: Dict<br/>labels: List of Tuples]
|
||
|
||
Store1 --> Batch2{More batches?}
|
||
Batch2 -->|Yes| NextBatch[Batch 2: Emails 21-40]
|
||
Batch2 -->|No| Consolidate
|
||
|
||
NextBatch --> Stats2[Same process<br/>15 total batches<br/>~20 seconds each]
|
||
Stats2 --> Batch2
|
||
|
||
Consolidate[Consolidate categories<br/>Merge duplicates<br/>Single LLM call<br/>~5 seconds]
|
||
|
||
Consolidate --> CacheSnap[Snap to cached categories<br/>Match against persistent cache<br/>~0.5 seconds]
|
||
|
||
CacheSnap --> Final[Final output<br/>10-12 categories<br/>300 labeled emails]
|
||
|
||
Final --> End([Labels ready for ML training])
|
||
|
||
style LLMCall1 fill:#ff6b6b
|
||
style Consolidate fill:#ff6b6b
|
||
style Stats2 fill:#ffd93d
|
||
style Final fill:#4ec9b0
|
||
</pre>
|
||
</div>
|
||
|
||
<h2>3. Key Discovery: Batched Labeling</h2>
|
||
|
||
<div class="code-section">
|
||
<strong>src/calibration/llm_analyzer.py:66-83</strong>
|
||
|
||
batch_size = 20 # NOT 1 email at a time!
|
||
|
||
for batch_idx in range(0, len(sample_emails), batch_size):
|
||
batch = sample_emails[batch_idx:batch_idx + batch_size]
|
||
|
||
# Single LLM call handles ENTIRE batch
|
||
batch_results = self._analyze_batch(batch, batch_idx)
|
||
|
||
# Returns BOTH categories AND labels for all 20 emails
|
||
for category, desc in batch_results.get('categories', {}).items():
|
||
discovered_categories[category] = desc
|
||
|
||
for email_id, category in batch_results.get('labels', []):
|
||
email_labels.append((email_id, category))
|
||
</div>
|
||
|
||
<div class="warning">
|
||
<h3>Why Batching Matters</h3>
|
||
<p><strong>Sequential (WRONG assumption):</strong> 300 emails × 3 sec/email = 900 seconds (15 minutes)</p>
|
||
<p><strong>Batched (ACTUAL):</strong> 15 batches × 20 sec/batch = 300 seconds (5 minutes)</p>
|
||
<p><strong>Savings:</strong> 10 minutes (67% faster than assumed)</p>
|
||
</div>
|
||
|
||
<h2>4. Single Batch Processing Detail</h2>
|
||
<div class="diagram">
|
||
<pre class="mermaid">
|
||
flowchart TD
|
||
Start([Batch of 20 emails]) --> Stats[Calculate Statistics<br/>~0.1 seconds]
|
||
|
||
Stats --> StatDetails[Domain analysis<br/>Recipient counts<br/>Attachment detection<br/>Keyword extraction]
|
||
|
||
StatDetails --> BuildList[Build email summaries<br/>For each email:<br/>ID + From + Subject + Preview]
|
||
|
||
BuildList --> Prompt[Construct LLM prompt<br/>~2KB text<br/>Contains:<br/>- Statistics summary<br/>- All 20 email summaries<br/>- Instructions<br/>- JSON schema]
|
||
|
||
Prompt --> LLM[LLM Call<br/>POST /api/generate<br/>qwen3:4b-instruct-2507-q8_0<br/>temp=0.1, max_tokens=2000<br/>~18-22 seconds]
|
||
|
||
LLM --> Response[LLM Response<br/>JSON with:<br/>categories: Dict<br/>labels: List of 20 Tuples]
|
||
|
||
Response --> Parse[Parse JSON<br/>Regex extraction<br/>Brace counting<br/>~0.05 seconds]
|
||
|
||
Parse --> Validate{Valid JSON?}
|
||
Validate -->|Yes| Extract[Extract data<br/>categories: 3-8 new<br/>labels: 20 tuples]
|
||
Validate -->|No| FallbackParse[Fallback parsing<br/>Try to salvage partial data]
|
||
|
||
FallbackParse --> Extract
|
||
|
||
Extract --> Return[Return batch results<br/>categories: Dict str→str<br/>labels: List Tuple str,str]
|
||
|
||
Return --> End([Merge with global results])
|
||
|
||
style LLM fill:#ff6b6b
|
||
style Parse fill:#4ec9b0
|
||
style FallbackParse fill:#ffd93d
|
||
</pre>
|
||
</div>
|
||
|
||
<h2>5. LLM Prompt Structure</h2>
|
||
|
||
<div class="code-section">
|
||
<strong>Actual prompt sent to LLM (src/calibration/llm_analyzer.py:196-232):</strong>
|
||
|
||
<no_think>You are analyzing emails to discover natural categories...
|
||
|
||
BATCH STATISTICS (20 emails):
|
||
- Top sender domains: example.com (5), company.org (3)...
|
||
- Avg recipients per email: 2.3
|
||
- Emails with attachments: 4/20
|
||
- Avg subject length: 42 chars
|
||
- Common keywords: meeting(3), report(2)...
|
||
|
||
EMAILS TO ANALYZE:
|
||
1. ID: maildir_allen-p__sent_mail_512
|
||
From: phillip.allen@enron.com
|
||
Subject: Re: AEC Volumes at OPAL
|
||
Preview: Here are the volumes...
|
||
|
||
2. ID: maildir_allen-p__sent_mail_513
|
||
From: phillip.allen@enron.com
|
||
Subject: Meeting Tomorrow
|
||
Preview: Can we schedule...
|
||
|
||
[... 18 more emails ...]
|
||
|
||
TASK:
|
||
1. Identify natural groupings based on PURPOSE
|
||
2. Create SHORT category names
|
||
3. Assign each email to exactly one category
|
||
4. CRITICAL: Copy EXACT email IDs
|
||
|
||
Return JSON:
|
||
{
|
||
"categories": {"Work": "daily business communication", ...},
|
||
"labels": [["maildir_allen-p__sent_mail_512", "Work"], ...]
|
||
}
|
||
</div>
|
||
|
||
<h2>6. Timing Breakdown - 300 Sample Emails</h2>
|
||
|
||
<table class="timing-table">
|
||
<tr>
|
||
<th>Operation</th>
|
||
<th>Per Batch (20 emails)</th>
|
||
<th>Total (15 batches)</th>
|
||
<th>% of Total Time</th>
|
||
</tr>
|
||
<tr>
|
||
<td>Calculate statistics</td>
|
||
<td>0.1 sec</td>
|
||
<td>1.5 sec</td>
|
||
<td>0.5%</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Build email summaries</td>
|
||
<td>0.05 sec</td>
|
||
<td>0.75 sec</td>
|
||
<td>0.2%</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Construct prompt</td>
|
||
<td>0.01 sec</td>
|
||
<td>0.15 sec</td>
|
||
<td>0.05%</td>
|
||
</tr>
|
||
<tr>
|
||
<td><strong>LLM API call</strong></td>
|
||
<td><strong>18-22 sec</strong></td>
|
||
<td><strong>270-330 sec</strong></td>
|
||
<td><strong>98%</strong></td>
|
||
</tr>
|
||
<tr>
|
||
<td>Parse JSON response</td>
|
||
<td>0.05 sec</td>
|
||
<td>0.75 sec</td>
|
||
<td>0.2%</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Merge results</td>
|
||
<td>0.02 sec</td>
|
||
<td>0.3 sec</td>
|
||
<td>0.1%</td>
|
||
</tr>
|
||
<tr>
|
||
<td colspan="2"><strong>SUBTOTAL: Batch Discovery</strong></td>
|
||
<td><strong>~300 seconds (5 min)</strong></td>
|
||
<td><strong>98.5%</strong></td>
|
||
</tr>
|
||
<tr>
|
||
<td colspan="2">Consolidation LLM call</td>
|
||
<td>5 seconds</td>
|
||
<td>1.3%</td>
|
||
</tr>
|
||
<tr>
|
||
<td colspan="2">Cache snapping (semantic matching)</td>
|
||
<td>0.5 seconds</td>
|
||
<td>0.2%</td>
|
||
</tr>
|
||
<tr>
|
||
<td colspan="2"><strong>TOTAL LABELING PHASE</strong></td>
|
||
<td><strong>~305 seconds (5 min)</strong></td>
|
||
<td><strong>100%</strong></td>
|
||
</tr>
|
||
</table>
|
||
|
||
<div class="warning">
|
||
<h3>Corrected Understanding</h3>
|
||
<p><strong>Original estimate:</strong> "~3 seconds per email" = 900 seconds for 300 emails</p>
|
||
<p><strong>Actual timing:</strong> ~20 seconds per batch of 20 = ~305 seconds for 300 emails</p>
|
||
<p><strong>Difference:</strong> 3× faster than original assumption</p>
|
||
<p><strong>Why:</strong> Batching allows LLM to see context across multiple emails and make better category decisions in a single inference pass.</p>
|
||
</div>
|
||
|
||
<h2>7. What Gets Created</h2>
|
||
|
||
<div class="diagram">
|
||
<pre class="mermaid">
|
||
flowchart LR
|
||
Input[300 sampled emails] --> Discovery[Category Discovery<br/>15 batches × 20 emails]
|
||
|
||
Discovery --> RawCats[Raw Categories<br/>~30-40 discovered<br/>May have duplicates:<br/>Work, work, Business, etc.]
|
||
|
||
RawCats --> Consolidate[Consolidation<br/>LLM merges similar<br/>~5 seconds]
|
||
|
||
Consolidate --> Merged[Merged Categories<br/>~12-15 categories<br/>Work, Financial, etc.]
|
||
|
||
Merged --> CacheSnap[Cache Snap<br/>Match against persistent cache<br/>~0.5 seconds]
|
||
|
||
CacheSnap --> Final[Final Categories<br/>10-12 categories]
|
||
|
||
Discovery --> RawLabels[Raw Labels<br/>300 tuples:<br/>email_id, category]
|
||
|
||
RawLabels --> UpdateLabels[Update label categories<br/>to match snapped names]
|
||
|
||
UpdateLabels --> FinalLabels[Final Labels<br/>300 training pairs]
|
||
|
||
Final --> Training[Training Data]
|
||
FinalLabels --> Training
|
||
|
||
Training --> MLTrain[Train LightGBM Model<br/>~5 seconds]
|
||
|
||
MLTrain --> Model[Trained Model<br/>1.8MB .pkl file]
|
||
|
||
style Discovery fill:#ff6b6b
|
||
style Consolidate fill:#ff6b6b
|
||
style Model fill:#4ec9b0
|
||
</pre>
|
||
</div>
|
||
|
||
<h2>8. Example Output</h2>
|
||
|
||
<div class="code-section">
|
||
<strong>discovered_categories (Dict[str, str]):</strong>
|
||
{
|
||
"Work": "daily business communication and coordination",
|
||
"Financial": "budgets, reports, financial planning",
|
||
"Meetings": "scheduling and meeting coordination",
|
||
"Technical": "system issues and technical discussions",
|
||
"Requests": "action items and requests for information",
|
||
"Reports": "status reports and summaries",
|
||
"Administrative": "HR, policies, company announcements",
|
||
"Urgent": "time-sensitive matters",
|
||
"Conversational": "casual check-ins and social",
|
||
"External": "communication with external partners"
|
||
}
|
||
|
||
<strong>sample_labels (List[Tuple[str, str]]):</strong>
|
||
[
|
||
("maildir_allen-p__sent_mail_1", "Financial"),
|
||
("maildir_allen-p__sent_mail_2", "Work"),
|
||
("maildir_allen-p__sent_mail_3", "Meetings"),
|
||
("maildir_allen-p__sent_mail_4", "Work"),
|
||
("maildir_allen-p__sent_mail_5", "Financial"),
|
||
... (300 total)
|
||
]
|
||
</div>
|
||
|
||
<h2>9. Why Batching is Critical</h2>
|
||
|
||
<table class="timing-table">
|
||
<tr>
|
||
<th>Approach</th>
|
||
<th>LLM Calls</th>
|
||
<th>Time/Call</th>
|
||
<th>Total Time</th>
|
||
<th>Quality</th>
|
||
</tr>
|
||
<tr>
|
||
<td><strong>Sequential (1 email/call)</strong></td>
|
||
<td>300</td>
|
||
<td>3 sec</td>
|
||
<td>900 sec (15 min)</td>
|
||
<td>Poor - no context</td>
|
||
</tr>
|
||
<tr>
|
||
<td><strong>Small batches (5 emails/call)</strong></td>
|
||
<td>60</td>
|
||
<td>8 sec</td>
|
||
<td>480 sec (8 min)</td>
|
||
<td>Fair - limited context</td>
|
||
</tr>
|
||
<tr>
|
||
<td><strong>Current (20 emails/call)</strong></td>
|
||
<td>15</td>
|
||
<td>20 sec</td>
|
||
<td>300 sec (5 min)</td>
|
||
<td>Good - sufficient context</td>
|
||
</tr>
|
||
<tr>
|
||
<td><strong>Large batches (50 emails/call)</strong></td>
|
||
<td>6</td>
|
||
<td>45 sec</td>
|
||
<td>270 sec (4.5 min)</td>
|
||
<td>Risk - may exceed token limits</td>
|
||
</tr>
|
||
</table>
|
||
|
||
<div class="warning">
|
||
<h3>Why 20 emails per batch?</h3>
|
||
<ul>
|
||
<li><strong>Token limit:</strong> 20 emails × ~150 tokens/email = ~3000 tokens input, well under 8K limit</li>
|
||
<li><strong>Context window:</strong> LLM can see patterns across multiple emails</li>
|
||
<li><strong>Speed:</strong> Minimizes API calls while staying within limits</li>
|
||
<li><strong>Quality:</strong> Enough examples to identify patterns, not so many that it gets confused</li>
|
||
</ul>
|
||
</div>
|
||
|
||
<h2>10. Configuration Parameters</h2>
|
||
|
||
<table class="timing-table">
|
||
<tr>
|
||
<th>Parameter</th>
|
||
<th>Location</th>
|
||
<th>Default</th>
|
||
<th>Effect on Timing</th>
|
||
</tr>
|
||
<tr>
|
||
<td>sample_size</td>
|
||
<td>CalibrationConfig</td>
|
||
<td>300</td>
|
||
<td>300 samples = 15 batches = 5 min</td>
|
||
</tr>
|
||
<tr>
|
||
<td>batch_size</td>
|
||
<td>llm_analyzer.py:62</td>
|
||
<td>20</td>
|
||
<td>Hardcoded - affects batch count</td>
|
||
</tr>
|
||
<tr>
|
||
<td>llm_batch_size</td>
|
||
<td>CalibrationConfig</td>
|
||
<td>50</td>
|
||
<td>NOT USED for discovery (misleading name)</td>
|
||
</tr>
|
||
<tr>
|
||
<td>temperature</td>
|
||
<td>LLM call</td>
|
||
<td>0.1</td>
|
||
<td>Lower = faster, more deterministic</td>
|
||
</tr>
|
||
<tr>
|
||
<td>max_tokens</td>
|
||
<td>LLM call</td>
|
||
<td>2000</td>
|
||
<td>Higher = potentially slower response</td>
|
||
</tr>
|
||
</table>
|
||
|
||
<h2>11. Full Calibration Timeline</h2>
|
||
|
||
<div class="diagram">
|
||
<pre class="mermaid">
|
||
gantt
|
||
title Calibration Phase Timeline (300 samples, 10k total emails)
|
||
dateFormat mm:ss
|
||
axisFormat %M:%S
|
||
|
||
section Sampling
|
||
Stratified sample (3% of 10k) :00:00, 01s
|
||
|
||
section Category Discovery
|
||
Batch 1 (emails 1-20) :00:01, 20s
|
||
Batch 2 (emails 21-40) :00:21, 20s
|
||
Batch 3 (emails 41-60) :00:41, 20s
|
||
Batch 4-13 (emails 61-260) :01:01, 200s
|
||
Batch 14 (emails 261-280) :04:21, 20s
|
||
Batch 15 (emails 281-300) :04:41, 20s
|
||
|
||
section Consolidation
|
||
LLM category merge :05:01, 05s
|
||
Cache snap :05:06, 00.5s
|
||
|
||
section ML Training
|
||
Feature extraction (300) :05:07, 06s
|
||
LightGBM training :05:13, 05s
|
||
Validation (100 emails) :05:18, 02s
|
||
Save model to disk :05:20, 00.5s
|
||
</pre>
|
||
</div>
|
||
|
||
<h2>12. Key Insights</h2>
|
||
|
||
<div class="critical">
|
||
<h3>1. Labels are NOT created sequentially</h3>
|
||
<p>The LLM creates labels as a byproduct of batch category discovery. There is NO separate "label each email one by one" phase.</p>
|
||
</div>
|
||
|
||
<div class="critical">
|
||
<h3>2. Batching is the optimization</h3>
|
||
<p>Processing 20 emails in a single LLM call (20 sec) is 3× faster than 20 individual calls (60 sec total).</p>
|
||
</div>
|
||
|
||
<div class="critical">
|
||
<h3>3. LLM time dominates everything</h3>
|
||
<p>98% of labeling phase time is LLM API calls. Everything else (parsing, merging, caching) is negligible.</p>
|
||
</div>
|
||
|
||
<div class="critical">
|
||
<h3>4. Consolidation is cheap</h3>
|
||
<p>Merging 30-40 raw categories into 10-12 final ones takes only ~5 seconds with a single LLM call.</p>
|
||
</div>
|
||
|
||
<h2>13. Optimization Opportunities</h2>
|
||
|
||
<table class="timing-table">
|
||
<tr>
|
||
<th>Optimization</th>
|
||
<th>Current</th>
|
||
<th>Potential</th>
|
||
<th>Tradeoff</th>
|
||
</tr>
|
||
<tr>
|
||
<td>Increase batch size</td>
|
||
<td>20 emails/batch</td>
|
||
<td>30-40 emails/batch</td>
|
||
<td>May hit token limits, slower per call</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Reduce sample size</td>
|
||
<td>300 samples (3%)</td>
|
||
<td>200 samples (2%)</td>
|
||
<td>Less training data, potentially worse model</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Parallel batching</td>
|
||
<td>Sequential 15 batches</td>
|
||
<td>3-5 concurrent batches</td>
|
||
<td>Requires async LLM client, more complex</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Skip consolidation</td>
|
||
<td>Always consolidate if >10 cats</td>
|
||
<td>Skip if <15 cats</td>
|
||
<td>May leave duplicate categories</td>
|
||
</tr>
|
||
<tr>
|
||
<td>Cache-first approach</td>
|
||
<td>Discover then snap to cache</td>
|
||
<td>Snap to cache, only discover new</td>
|
||
<td>Less adaptive to new mailbox types</td>
|
||
</tr>
|
||
</table>
|
||
|
||
<script>
|
||
mermaid.initialize({
|
||
startOnLoad: true,
|
||
theme: 'default',
|
||
flowchart: {
|
||
useMaxWidth: true,
|
||
htmlLabels: true,
|
||
curve: 'basis'
|
||
},
|
||
gantt: {
|
||
useWidth: 1200
|
||
}
|
||
});
|
||
</script>
|
||
</body>
|
||
</html>
|