system-prompts-and-models-of-ai-tools/dealix/docs/ops/DEPLOY_NOW.md

Dealix Lead Machine — Deploy Now

This bundle contains the full Provider Adapters + Data Lake + Lead Graph + Outreach Prep build, ready to land on main in two commits (73b5bb7 + 7aa2302). Both are inside the single patch file.

Patch dry-applied to a fresh clone of VoXc2/dealix and 32 unit tests passed — so git am will work on your local repo if your main is at 2a9df98 or earlier.

What's in the bundle

File	What
0001-feat-lead-machine.patch	Single git patch — apply with git am
dealix-lead-machine-files.tar.gz	Same files as a tarball — copy into your repo if git am fails
DEPLOY_NOW.md	This doc

Commit content:

• 5 provider chains (Search / Maps / Crawler / Tech / EmailIntel) with env-gated fallbacks
• 7 new SQLAlchemy tables (raw_lead_imports, raw_lead_rows, accounts, contacts, signals, lead_scores, data_suppression_list)
• 4 normalize/dedupe/score/enrichment pipelines
• 19 new API endpoints (/leads/discover/local, /leads/discover/web, /leads/enrich/full, /leads/enrich/batch, all /data/*, all /outreach/*)
• 4 CLI scripts (import_leads, audit_lead_file, discover_local_to_csv, export_outreach_ready)
• 4 architecture docs
• 32 smoke tests passing
• Updated .env.example with all new env vars
• Updated /api/v1/prospect/search-diag with tier1_ready / tier2_ready flags

---

Step 1 — Apply the patch

On your laptop, in your dealix repo:

cd /path/to/dealix
git checkout main
git pull --rebase origin main

# Apply the patch
git am < /path/to/0001-feat-lead-machine.patch

# OR if git am has trouble, extract the tarball and stage manually:
# tar xzf /path/to/dealix-lead-machine-files.tar.gz
# git add -A
# git commit -m "feat(lead-machine): provider chains + Data Lake + Lead Graph"

Verify:

git log --oneline -1
# should show: feat(lead-machine): provider chains + Data Lake + Lead Graph + outreach prep

Step 2 — Push

If you don't have GitHub CLI auth set up:

gh auth login
# choose: GitHub.com → HTTPS → Login with browser

Then:

git push origin main

Do NOT reuse the PAT you pasted in chat. It should be revoked.

Step 3 — Add Railway env vars

Railway → project → environment Dealix → service web → Variables. Add:

# ── Layer 1 (required) ──
DATABASE_URL=${{Postgres.DATABASE_URL}}
GOOGLE_SEARCH_API_KEY=<your_key>
GOOGLE_SEARCH_CX=75ae2277dfd754a1a
GROQ_API_KEY=<your_key>
SENTRY_DSN=<your_dsn>

# ── Layer 2 (lead engine) ──
GOOGLE_MAPS_API_KEY=<your_places_key>

Click Review → Deploy.

For the Google Maps key:

1. Google Cloud Console → APIs & Services → Library → enable Places API.
2. Credentials → Create Credentials → API key.
3. Restrict the key: API restrictions → Places API only.
4. Paste into Railway as GOOGLE_MAPS_API_KEY.

Step 4 — Verify deploy

After Railway shows the deploy as Active:

# Tier-readiness (one-shot check)
curl https://api.dealix.me/api/v1/prospect/search-diag | jq '{tier1_ready, tier2_ready, hint}'

# Should show tier1_ready: true, tier2_ready: true (after Maps key added)

# List Saudi industries + cities
curl https://api.dealix.me/api/v1/leads/discover/local-industries | jq '.industries[0:3], .cities[0:3]'

# Pull 20 dental clinics in Riyadh
curl -X POST https://api.dealix.me/api/v1/leads/discover/local \
  -H 'content-type: application/json' \
  -d '{"industry":"dental_clinic","city":"riyadh","max_results":20,"hydrate_details":true}' | jq '.data.results[0:3]'

# Single-account enrichment
curl -X POST https://api.dealix.me/api/v1/leads/enrich/full \
  -H 'content-type: application/json' \
  -d '{"company_name":"Foodics","domain":"foodics.com","city":"Riyadh","sector":"saas","level":"standard"}' | jq '.score, .data_quality, .providers_used'

Step 5 — First end-to-end run (your data → outreach queue)

# 1. Audit a CSV before uploading (offline)
python scripts/audit_lead_file.py vendor_file.csv

# 2. Upload + run full pipeline (normalize → dedupe → enrich)
python scripts/import_leads.py vendor_file.csv \
    --source-name "vendor_x_2026_q2" \
    --source-type paid \
    --allowed-use "business_contact_research_only" \
    --risk-level high \
    --auto-pipeline

# 3. Get outreach-ready leads
python scripts/export_outreach_ready.py --priority P0,P1 --max 50 --out today_outreach_50.csv

Or skip the file and pull live from Maps:

python scripts/discover_local_to_csv.py dental_clinic riyadh --max 20
python scripts/import_leads.py dental_clinic_riyadh.csv \
    --source-name "maps_dental_clinic_riyadh" \
    --source-type google_maps --auto-pipeline
python scripts/export_outreach_ready.py --priority P0,P1 --max 20 --out riyadh_dentists_outreach.csv

Step 6 — Check the local smoke tests pass

cd /path/to/dealix
python -m pytest tests/unit/test_pipelines_smoke.py tests/unit/test_provider_smoke.py -q -o "addopts="
# expected: 32 passed, 2 skipped (live provider tests skip without keys)

---

Compliance reminders (already enforced in code)

• Every imported row carries source_type, allowed_use, consent_status, risk_level.
• /outreach/prepare-from-data checks data_suppression_list + opt_out before producing ready entries.
• approval_required=True is the default for all queued messages — Sami still approves.
• WhatsApp is inbound only in Dealix. Don't add an outbound WhatsApp adapter.
• LinkedIn is manual research + manual send only. Don't add scraping.

Troubleshooting

Problem	Likely cause	Fix
tier1_ready: false after deploy	One of DATABASE_URL/GOOGLE_SEARCH_API_KEY+CX/LLM/SENTRY_DSN missing	Check search-diag output, add the missing one, re-deploy
/leads/discover/local returns status: no_key	GOOGLE_MAPS_API_KEY not set or empty in Railway	Add it, click Review → Deploy
/data/import returns skipped_db_unreachable	DB is the same problem as everywhere — DATABASE_URL empty	Set DATABASE_URL=${{Postgres.DATABASE_URL}} (with the literal Railway template syntax)
git am fails	local uncommitted work	git stash first, then git am, then git stash pop
401 from Google Places	billing not enabled or wrong restriction	GCP Console → enable billing on the project, restrict key to Places API