Deltek Vantagepoint Release Notes Explorer

A self-updating static website that lets you search and browse all Deltek Vantagepoint release notes — defects, enhancements, regulatory changes, and security updates — across every version from 2.0 onward. Runs entirely in the browser using SQLite compiled to WebAssembly. No server required after the initial download.

Live demo: deploy to GitHub Pages (see below).

---

Features

🔍 Full-text search	Debounced, non-blocking — queries run in a Web Worker
🏷️ Filter by type	Defects · Enhancements · Regulatory · Ported Defects
🗂️ Application Area tree	Three-level drill-down matching the in-app navigation breadcrumbs
🔀 Ported defect detection	Same defect number across multiple versions → "⇢ N versions" badge
📦 Filter by version	Drill into any patch release
📄 Paginated results	250 rows/page — no DOM storm on large result sets
⬇️ Streaming download	Progress bar during initial DB load, main thread never blocked
🌑 Dark theme

Application Area hierarchy

Breadcrumbs are parsed directly from the release notes pages and stored as three nav levels:

Level	Example
nav_level1	Hubs · Billing · My Stuff · Transaction Center
nav_level2	Projects · Interactive Billing · Reporting
nav_level3	Planning · AP Vouchers

Regulatory and Security issues are grouped under synthetic top-level nodes (Regulatory, Security) so that jurisdiction names (Federal, California, Payroll) don't pollute the Application Area sidebar alongside app module names.

Application Area
├── Billing
│   ├── Batch Billing
│   └── Interactive Billing
├── Hubs
│   └── Projects
│       ├── Planning
│       └── Project
├── Regulatory          ← all regulatory items grouped here
│   ├── Federal
│   └── Payroll
│       ├── California
│       └── Federal
└── Security            ← all security items grouped here
    └── API

---

Quick start

Run locally (demo data — no scraping needed)

git clone https://github.com/YOUR_USER/vantagepoint-release-notes.git
cd vantagepoint-release-notes

pip install -r scraper/requirements.txt
python scraper/seed_demo.py       # populates db/release_notes.db with sample data
cp db/release_notes.db web/

python serve.py                   # http://localhost:8001  (opens browser automatically)

Why serve.py instead of python -m http.server? Python's built-in HTTP server does not handle Range: requests — the progress bar works but shows no percentage. serve.py adds proper 206 Partial Content support so the progress bar is accurate and local dev matches GitHub Pages behaviour exactly.

python serve.py 8080              # custom port
python serve.py 8001 /path/to/web # custom port + directory

Scrape all real data

# Full fresh scrape (all versions, ~30–60 min, ~200+ pages)
python scraper/scraper.py

# Resume — skip old releases, always re-check last 10 (detects retroactive amendments)
python scraper/scraper.py --resume --recheck 10

# Only specific major versions
python scraper/scraper.py --versions 7.0 2025.1

# Preview without writing to DB
python scraper/scraper.py --dry-run --versions 2025.1

# All options
python scraper/scraper.py --help

After scraping:

cp db/release_notes.db web/
python serve.py

---

Deploy to GitHub Pages

One-time setup

1. Fork / push this repo to GitHub
2. Settings → Pages → Source: set to "GitHub Actions"
3. Settings → Actions → General → Workflow permissions: "Read and write permissions"

The included workflow (.github/workflows/scrape-deploy.yml) will:

• Run the parser unit tests first — deploy only if they pass
• Scrape with --resume --recheck 10 --concurrency 4 (fast on subsequent runs)
• Deploy web/ to the gh-pages branch
• Repeat every Monday at 06:00 UTC

Manual trigger

Go to Actions → Scrape & Deploy Release Notes → Run workflow.

---

Project structure

├── scraper/
│   ├── scraper.py          Main scraper — fetch, parse, persist
│   ├── seed_demo.py        Seed a working demo DB without network access
│   └── requirements.txt    requests, beautifulsoup4, lxml, pytest
├── tests/
│   └── test_parser.py      30 offline unit tests for parse_html and utilities
├── web/
│   ├── index.html          Single-file static site (sql.js via Web Worker)
│   └── release_notes.db    Generated SQLite database (committed for GH Pages)
├── serve.py                Local dev server with HTTP Range request support
└── .github/workflows/
    └── scrape-deploy.yml   CI: test → scrape → deploy

---

How the scraper works

Version discovery

The scraper fetches the Deltek master index at:

https://help.deltek.com/product/Vantagepoint/ReleaseNotes/

All version-specific index URLs are extracted from this page automatically. New versions (e.g. 2026.3) are picked up on the next weekly run with no code changes. A hardcoded fallback list is used if the master page is unreachable.

Resume and change detection

All releases sorted newest-first
├── Last N releases (--recheck N, default 5) → always re-fetched
│   ├── content_hash unchanged → update timestamp only, skip DB write
│   └── content_hash changed  → re-upsert all issues, record diff
│       ├── added_keys   (new defect numbers on the page)
│       ├── removed_keys (defect numbers no longer listed)
│       └── modified_keys (same defect, description text amended)
└── Older releases → skipped entirely

Every run appends a row to scrape_history so you have a complete audit trail of when each page changed and what changed in it.

Database schema

releases        -- one row per patch release (7.0.11, 2025.1.4, …)
issues          -- one row per unique issue (defect, enhancement, regulatory, security)
                -- issue_key: stable SHA-256 identity, excludes description
                -- first_seen_at / updated_at: timestamps
issue_versions  -- many-to-many: which releases contain which issue (ported defects)
scrape_log      -- latest scrape state per URL (content_hash, page_last_updated)
scrape_history  -- append-only audit: every scrape attempt, diffs as JSON arrays
issues_fts      -- FTS5 virtual table for full-text search

Parser design

parse_html(html: str) is a pure function decoupled from HTTP — pass any HTML string to it. This makes it fully unit-testable without network access.

Key parsing decisions:

Problem	Solution
Defect1541790 (no space) consumed as breadcrumb	DEFECT_RE uses \s* not \s+
Enhancement titles in <p><strong> mistaken for module names	is_heading_only AND len ≤ 25 chars threshold
Enhancement title and description were emitted as two issues	pending_title buffer — bold title held until next plain paragraph
Regulatory jurisdiction names (Federal, California) pollute Application Area	_nav_for_section() remaps to Regulatory >> Federal etc.
Same defect number appearing twice on one page	Deduplication uses make_issue_key() (same hash as DB)

---

Running tests

# No pytest needed
python tests/test_parser.py

# With pytest
pip install pytest
pytest tests/ -v

Tests are offline — no network access required. The 30 tests cover:

• DEFECT_RE matching (standard and no-space variants)
• Breadcrumb parsing at all three levels
• Enhancement title + description merging
• Regulatory/Security nav remapping
• _nav_for_section() all section types
• make_issue_key stability across description changes
• make_content_hash change detection
• _diff add/remove/modify classification
• Deduplication
• page_last_updated extraction

---

Notes

• Not affiliated with Deltek. The underlying release note content is © Deltek Inc.
• The scraper uses a 0.5 s delay between requests per worker. Please be respectful of Deltek's servers.