import os, requests, time, base64
from collections import defaultdict
GH_KEY = os.environ["GREENHOUSE_API_KEY"]
MV = os.environ["MAVERA_API_KEY"]
GH_BASE = "https://harvest.greenhouse.io/v1"
MV_BASE = "https://app.mavera.io/api/v1"
MV_H = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}
gh_auth = base64.b64encode(f"{GH_KEY}:".encode()).decode()
GH_H = {"Authorization": f"Basic {gh_auth}"}
SOURCES = ["LinkedIn", "Referral", "Careers Page"]
OUTCOMES = ["hired", "rejected"]
def gh_get(path, params=None):
r = requests.get(f"{GH_BASE}{path}", headers=GH_H, params=params or {})
if r.status_code == 429:
time.sleep(10)
return gh_get(path, params)
r.raise_for_status()
return r.json()
# 1. Pull candidates and their applications
candidates = []
page = 1
while len(candidates) < 500:
batch = gh_get("/candidates", {"per_page": 100, "page": page})
if not batch:
break
candidates.extend(batch)
page += 1
# 2. Group by source × outcome
segments = defaultdict(list)
for c in candidates:
for app in c.get("applications", []):
source_name = (app.get("source", {}) or {}).get("public_name", "Unknown")
status = app.get("status", "active")
if source_name in SOURCES and status in OUTCOMES:
segments[(source_name, status)].append({
"name": f"{c.get('first_name','')} {c.get('last_name','')}",
"title": c.get("title", ""),
"company": c.get("company", ""),
"application_count": len(c.get("applications", [])),
})
# 3. Create Mavera personas per segment
persona_ids = []
for (source, outcome), members in segments.items():
if len(members) < 3:
continue
titles = list({m["title"] for m in members if m["title"]})[:5]
companies = list({m["company"] for m in members if m["company"]})[:5]
p = requests.post(f"{MV_BASE}/personas", headers=MV_H, json={
"name": f"GH: {source} → {outcome.title()}",
"description": (
f"Candidates from {source} who were {outcome}. "
f"N={len(members)}. Titles: {', '.join(titles[:3])}. "
f"Companies: {', '.join(companies[:3])}."
),
"demographic": {"job_titles": titles},
"psychographic": {
"source": source,
"outcome": outcome,
"mindset": f"Candidate who was {outcome} via {source}",
},
}).json()
persona_ids.append({"id": p["id"], "label": f"{source} → {outcome.title()}"})
print(f"Persona: {p['id']} — {source} → {outcome.title()} ({len(members)} candidates)")
time.sleep(0.3)
# 4. Run Focus Group with employer messaging
EMPLOYER_MESSAGE = """Join a team that ships fast, learns faster, and celebrates wins together.
We offer competitive comp, unlimited PTO, and a culture where engineers own their roadmap.
"Best decision I ever made." — Senior Engineer, 2 years"""
fg = requests.post(f"{MV_BASE}/focus-groups", headers=MV_H, json={
"name": "Employer Brand Messaging Test",
"persona_ids": [p["id"] for p in persona_ids],
"questions": [
"How authentic does this employer message feel on a scale of 1-10?",
"Would this message make you more or less likely to apply? Why?",
"What specific claim feels most credible? Least credible?",
"How would you describe this company's culture to a friend based on this message?",
],
"context": EMPLOYER_MESSAGE,
"responses_per_persona": 3,
}).json()
# 5. Poll for results
for _ in range(20):
time.sleep(5)
data = requests.get(f"{MV_BASE}/focus-groups/{fg['id']}", headers=MV_H).json()
if data.get("status") == "completed":
break
for resp in data.get("responses", []):
label = next((p["label"] for p in persona_ids if p["id"] == resp.get("persona_id")), "?")
print(f"\n[{label}] {resp.get('question','')[:60]}")
print(f" → {resp.get('answer','')[:250]}")
