Skip to main content

Knowledge > Processes > Nightly Google Drive Sync

Nightly Google Drive Sync

How knowledge/ canonical files sync to Google Drive so the CEO can browse them at C:\Users\johnm\ChurchWiseAI\ via Google Drive for Desktop.

Overview​

The knowledge system uses a derive pipeline (pnpm derive) that reads canonical YAML data and markdown narratives, then propagates them to downstream targets. One target type is gdrive:, which converts source files to Google Docs-compatible HTML and uploads them to the corresponding Google Drive folder.

Manifest Configuration​

knowledge/manifest.yaml declares which sources have Google Drive targets. Each gdrive: target specifies a destination folder path.

manifest.yaml entries with type: nightly

  data/pricing.yaml    → gdrive:03-Strategy/Pricing/
  data/features.yaml   → gdrive:02-Products/
  data/products.yaml   → gdrive:02-Products/
  data/policies.yaml   → gdrive:01-Business/Legal/
  narrative/vision.md  → gdrive:01-Business/Company/
  narrative/competitive.md  → gdrive:03-Strategy/Competitive/
  narrative/sales-playbook.md → gdrive:05-Sales/
  narrative/brand.md   → gdrive:04-Marketing/Brand/
  narrative/strategy.md → gdrive:03-Strategy/Roadmap/
  narrative/customer-journey.md → gdrive:05-Sales/
  narrative/operations.md → gdrive:06-Operations/SOPs/

Pipeline Steps​

1. READ SOURCE FILE
   source = read knowledge/data/*.yaml or knowledge/narrative/*.md

2. DETECT FORMAT
   IF source ends with .yaml:
       data = parse YAML
       SWITCH on source filename:
           pricing.yaml  → html = pricingYamlToHtml(data, metadata)
           features.yaml → html = featuresYamlToHtml(data, metadata)
           products.yaml → html = productsYamlToHtml(data, metadata)
           policies.yaml → html = policiesYamlToHtml(data, metadata)
   ELSE IF source ends with .md:
       html = markdownToGDocHtml(markdown_content, metadata)

3. CONVERT TO GOOGLE DOCS HTML
   All converters produce clean HTML suitable for Google Docs import:
     - Headings: <h1>, <h2>, <h3>
     - Tables: <table> with <th>/<td>
     - Lists: <ul>/<ol> with <li>
     - Inline: <strong>, <em>, <code>, <a>
     - Code blocks: <pre>
     - Links preserved as clickable <a href>

4. ADD METADATA FOOTER
   metadataFooter(metadata) appends:
     <hr>
     <p style="color: #999; font-size: 0.8em;">
       Generated from {source_file} · {date} · Commit: {hash}
     </p>

5. UPLOAD TO GOOGLE DRIVE
   target folder = gdrive:{folder_path}
   Maps to: Google Drive > ChurchWiseAI > {folder_path}
   Synced to local via Drive for Desktop:
     C:\Users\johnm\ChurchWiseAI\{folder_path}

Converter Functions (derive-gdrive.ts)​

markdownToGDocHtml(markdown, metadata?)​

Converts markdown narratives (vision, brand, strategy, etc.) to HTML.

HANDLES:
  # H1, ## H2, ### H3 headings
  **bold**, *italic*, `code`
  [link text](url)
  - unordered lists, 1. ordered lists
  | table | rows |
  ``` code blocks ```
  Blank lines → paragraph breaks

INLINE PROCESSING ORDER:
  1. Code spans: `text` → <code>text</code>
  2. Links: [text](url) → <a href="url">text</a>
  3. Bold: **text** → <strong>text</strong>  (before italic)
  4. Italic: *text* → <em>text</em>

STATE MACHINE tracks:
  inCodeBlock, inTable, inUl, inOl, paragraphLines[]
  Flushes accumulated state on context switches

pricingYamlToHtml(pricing, metadata?)​

Converts data/pricing.yaml to a formatted pricing guide.

OUTPUT:
  <h1>ChurchWiseAI Pricing Guide</h1>
  <h2>Products & Pricing</h2>
  <table> product | monthly | annual | one-time </table>
  <h2>Billing Rules</h2>
  <ul> key-value list from billing_rules </ul>
  <h2>Free Trials</h2>
  <p> list of trial-enabled products with durations </p>

featuresYamlToHtml(features, metadata?)​

Converts data/features.yaml to a feature matrix.

OUTPUT:
  <h1>ChurchWiseAI Feature Matrix</h1>
  <h2>Tier Overview</h2>
  <table> feature | starter | pro | suite </table>
  <h2>Features by Tier</h2>
  Per-tier bulleted feature lists
  <h2>Add-Ons</h2>
  Named add-ons with descriptions and feature bullets

productsYamlToHtml(products, metadata?)​

Converts data/products.yaml to product portfolio docs.

OUTPUT:
  Company info, brand info
  Products summary table: name | tagline | URL | status
  Product detail sections: description, what_it_sells, codebase

policiesYamlToHtml(policies, metadata?)​

Converts data/policies.yaml to formatted policy summary.

OUTPUT:
  Cancellation policy, billing rules, refund policy
  Price change policy, trial policy (chat vs voice)
  Data handling section (nested key-value lists)

Drive Folder Structure​

C:\Users\johnm\ChurchWiseAI\    (Google Drive for Desktop sync)
├── 01-Business/
│   ├── Company/       ← narrative/vision.md
│   └── Legal/         ← data/policies.yaml
├── 02-Products/       ← data/features.yaml, data/products.yaml
├── 03-Strategy/
│   ├── Competitive/   ← narrative/competitive.md
│   ├── Pricing/       ← data/pricing.yaml
│   └── Roadmap/       ← narrative/strategy.md
├── 04-Marketing/
│   └── Brand/         ← narrative/brand.md
├── 05-Sales/          ← narrative/sales-playbook.md, narrative/customer-journey.md
└── 06-Operations/
    └── SOPs/          ← narrative/operations.md

Running the Sync​

FULL PIPELINE (all sources):
  cd C:\dev\knowledge && pnpm derive --all

SINGLE SOURCE:
  cd C:\dev\knowledge && pnpm derive narrative/vision.md

DRY RUN (preview only):
  cd C:\dev\knowledge && pnpm derive data/pricing.yaml --dry-run

DRIFT CHECK (CI gate):
  cd C:\dev\knowledge && pnpm derive --check
  EXIT 0 = clean, EXIT 1 = drift detected

The derive pipeline also handles non-gdrive targets (regenerate pricing.ts, verify marketing pages, upsert product_knowledge), but those are documented in processes/system/derive-pipeline.md.