Floke
199a09f514
- Added searchStrategyICP, digitalSignals, and targetPages to the JSON output in market_intel_orchestrator.py. - Updated the prompt and fallback error handling to include the new fields.
660 lines
28 KiB
Python
660 lines
28 KiB
Python
import argparse
|
|
import json
|
|
import os
|
|
import sys # Import sys for stderr
|
|
import requests
|
|
from bs4 import BeautifulSoup
|
|
import logging
|
|
from datetime import datetime
|
|
import re # Für Regex-Operationen
|
|
|
|
# --- AUTARKES LOGGING SETUP --- #
|
|
def create_self_contained_log_filename(mode):
|
|
"""
|
|
Erstellt einen zeitgestempelten Logdateinamen für den Orchestrator.
|
|
Verwendet ein festes Log-Verzeichnis innerhalb des Docker-Containers.
|
|
NEU: Nur eine Datei pro Tag, um Log-Spam zu verhindern.
|
|
"""
|
|
log_dir_path = "/app/Log" # Festes Verzeichnis im Container
|
|
if not os.path.exists(log_dir_path):
|
|
os.makedirs(log_dir_path, exist_ok=True)
|
|
|
|
# Nur Datum verwenden, nicht Uhrzeit, damit alle Runs des Tages in einer Datei landen
|
|
date_str = datetime.now().strftime("%Y-%m-%d")
|
|
filename = f"{date_str}_market_intel.log"
|
|
return os.path.join(log_dir_path, filename)
|
|
|
|
log_filename = create_self_contained_log_filename("market_intel_orchestrator")
|
|
logging.basicConfig(
|
|
level=logging.DEBUG,
|
|
format='[%(asctime)s] %(levelname)s [%(funcName)s]: %(message)s',
|
|
datefmt='%Y-%m-%d %H:%M:%S',
|
|
handlers=[
|
|
logging.FileHandler(log_filename, mode='a', encoding='utf-8'),
|
|
logging.StreamHandler(sys.stderr)
|
|
]
|
|
)
|
|
logger = logging.getLogger(__name__)
|
|
# --- END AUTARKES LOGGING SETUP --- #
|
|
|
|
def load_gemini_api_key(file_path="gemini_api_key.txt"):
|
|
try:
|
|
with open(file_path, "r") as f:
|
|
api_key = f.read().strip()
|
|
return api_key
|
|
except Exception as e:
|
|
logger.critical(f"Fehler beim Laden des Gemini API Keys: {e}")
|
|
raise
|
|
|
|
def load_serp_api_key(file_path="serpapikey.txt"):
|
|
"""Lädt den SerpAPI Key. Gibt None zurück, wenn nicht gefunden."""
|
|
try:
|
|
if os.path.exists(file_path):
|
|
with open(file_path, "r") as f:
|
|
return f.read().strip()
|
|
# Fallback: Versuche Umgebungsvariable
|
|
return os.environ.get("SERP_API_KEY")
|
|
except Exception as e:
|
|
logger.warning(f"Konnte SerpAPI Key nicht laden: {e}")
|
|
return None
|
|
|
|
def get_website_text(url):
|
|
logger.info(f"Scraping URL: {url}")
|
|
try:
|
|
headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'}
|
|
response = requests.get(url, headers=headers, timeout=10)
|
|
response.raise_for_status()
|
|
soup = BeautifulSoup(response.text, 'lxml')
|
|
for tag in soup(['script', 'style', 'nav', 'footer', 'header']):
|
|
tag.decompose()
|
|
text = soup.get_text(separator=' ', strip=True)
|
|
# Bereinigung des Textes von nicht-druckbaren Zeichen
|
|
text = re.sub(r'[^\x20-\x7E\n\r\t]', '', text)
|
|
return text[:10000] # Limit für besseren Kontext
|
|
except Exception as e:
|
|
logger.error(f"Scraping failed for {url}: {e}")
|
|
return None
|
|
|
|
def serp_search(query, num_results=3):
|
|
"""Führt eine Google-Suche über SerpAPI durch."""
|
|
api_key = load_serp_api_key()
|
|
if not api_key:
|
|
logger.warning("SerpAPI Key fehlt. Suche übersprungen.")
|
|
return []
|
|
|
|
logger.info(f"SerpAPI Suche: {query}")
|
|
try:
|
|
params = {
|
|
"engine": "google",
|
|
"q": query,
|
|
"api_key": api_key,
|
|
"num": num_results,
|
|
"hl": "de",
|
|
"gl": "de"
|
|
}
|
|
response = requests.get("https://serpapi.com/search", params=params, timeout=20)
|
|
response.raise_for_status()
|
|
data = response.json()
|
|
|
|
results = []
|
|
if "organic_results" in data:
|
|
for result in data["organic_results"]:
|
|
results.append({
|
|
"title": result.get("title"),
|
|
"link": result.get("link"),
|
|
"snippet": result.get("snippet")
|
|
})
|
|
return results
|
|
except Exception as e:
|
|
logger.error(f"SerpAPI Fehler: {e}")
|
|
return []
|
|
|
|
def _extract_target_industries_from_context(context_content):
|
|
md = context_content
|
|
# Versuche verschiedene Muster für die Tabelle, falls das Format variiert
|
|
step2_match = re.search(r'##\s*Schritt\s*2:[\s\S]*?(?=\n##\s*Schritt\s*\d:|\s*$)', md, re.IGNORECASE)
|
|
if not step2_match:
|
|
# Fallback: Suche nach "Zielbranche" irgendwo im Text
|
|
match = re.search(r'Zielbranche\s*\|?\s*([^|\n]+)', md, re.IGNORECASE)
|
|
if match:
|
|
return [s.strip() for s in match.group(1).split(',')]
|
|
return []
|
|
|
|
table_lines = []
|
|
in_table = False
|
|
for line in step2_match.group(0).split('\n'):
|
|
if line.strip().startswith('|'):
|
|
in_table = True
|
|
table_lines.append(line.strip())
|
|
elif in_table: break
|
|
|
|
if len(table_lines) < 3: return []
|
|
header = [s.strip() for s in table_lines[0].split('|') if s.strip()]
|
|
industry_col = next((h for h in header if re.search(r'zielbranche|segment|branche|industrie', h, re.IGNORECASE)), None)
|
|
if not industry_col: return []
|
|
|
|
col_idx = header.index(industry_col)
|
|
industries = []
|
|
for line in table_lines[2:]:
|
|
cells = [s.strip() for s in line.split('|') if s.strip()]
|
|
if len(cells) > col_idx: industries.append(cells[col_idx])
|
|
return list(set(industries))
|
|
|
|
def _extract_json_from_text(text):
|
|
"""
|
|
Versucht, ein JSON-Objekt aus einem Textstring zu extrahieren,
|
|
unabhängig von Markdown-Formatierung (```json ... ```).
|
|
"""
|
|
try:
|
|
# 1. Versuch: Direktersatz von Markdown-Tags (falls vorhanden)
|
|
clean_text = text.replace("```json", "").replace("```", "").strip()
|
|
return json.loads(clean_text)
|
|
except json.JSONDecodeError:
|
|
pass
|
|
|
|
try:
|
|
# 2. Versuch: Regex Suche nach dem ersten { und letzten }
|
|
json_match = re.search(r"(\{[\s\S]*\})", text)
|
|
if json_match:
|
|
return json.loads(json_match.group(1))
|
|
except json.JSONDecodeError:
|
|
pass
|
|
|
|
logger.error(f"JSON Parsing fehlgeschlagen. Roher Text: {text[:500]}...")
|
|
return None
|
|
|
|
def generate_search_strategy(reference_url, context_content):
|
|
logger.info(f"Generating strategy for {reference_url}")
|
|
api_key = load_gemini_api_key()
|
|
target_industries = _extract_target_industries_from_context(context_content)
|
|
homepage_text = get_website_text(reference_url)
|
|
|
|
# Switch to stable 2.5-pro model (which works for v1beta)
|
|
GEMINI_API_URL = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-pro:generateContent?key={api_key}"
|
|
|
|
prompt = f"""
|
|
You are a B2B Market Intelligence Architect.
|
|
|
|
--- STRATEGIC CONTEXT ---
|
|
{context_content}
|
|
|
|
--- EXTRACTED TARGET INDUSTRIES ---
|
|
{', '.join(target_industries)}
|
|
|
|
--- REFERENCE CLIENT HOMEPAGE ---
|
|
{homepage_text[:10000] if homepage_text else "No Homepage Text"}
|
|
|
|
--- TASK ---
|
|
Based on the context and the reference client's homepage, develop a search strategy to find similar companies (competitors/lookalikes) and audit them to find sales triggers.
|
|
|
|
1. **summaryOfOffer**: A 1-sentence summary of what the reference client sells.
|
|
2. **idealCustomerProfile**: A concise definition of the Ideal Customer Profile (ICP) based on the reference client.
|
|
3. **searchStrategyICP**: A detailed description of the Ideal Customer Profile (ICP) based on the analysis.
|
|
4. **digitalSignals**: Identification and description of relevant digital signals that indicate purchase interest or engagement.
|
|
5. **targetPages**: A list of the most important target pages on the company website relevant for marketing and sales activities.
|
|
6. **signals**: Identify exactly 4 specific digital signals.
|
|
- **CRITICAL**: One signal MUST be "Technographic / Incumbent Search". It must look for existing competitor software or legacy systems that our offer replaces or complements (e.g., "Uses SAP Ariba", "Has Supplier Portal", "Uses Salesforce").
|
|
- The other 3 signals should focus on business pains or strategic fit (e.g., "Sustainability Report", "Supply Chain Complexity").
|
|
|
|
--- SIGNAL DEFINITION ---
|
|
For EACH signal, you MUST provide:
|
|
- `id`: A unique ID (e.g., "sig_1").
|
|
- `name`: A short, descriptive name.
|
|
- `description`: What does this signal indicate?
|
|
- `targetPageKeywords`: A list of 3-5 keywords to look for on a company's website (e.g., ["career", "jobs"] for a hiring signal).
|
|
- `proofStrategy`: An object containing:
|
|
- `likelySource`: Where on the website or web is this info found? (e.g., "Careers Page").
|
|
- `searchQueryTemplate`: A Google search query to find this. Use `{{COMPANY}}` as a placeholder for the company name.
|
|
Example: `site:{{COMPANY}} "software engineer" OR "developer"`
|
|
|
|
--- OUTPUT FORMAT ---
|
|
Return ONLY a valid JSON object.
|
|
{{
|
|
"summaryOfOffer": "...",
|
|
"idealCustomerProfile": "...",
|
|
"searchStrategyICP": "...",
|
|
"digitalSignals": "...",
|
|
"targetPages": "...",
|
|
"signals": [
|
|
{{
|
|
"id": "sig_1",
|
|
"name": "...",
|
|
"description": "...",
|
|
"targetPageKeywords": ["..."],
|
|
"proofStrategy": {{
|
|
"likelySource": "...",
|
|
"searchQueryTemplate": "..."
|
|
}}
|
|
}},
|
|
...
|
|
]
|
|
}}
|
|
"""
|
|
|
|
payload = {"contents": [{"parts": [{"text": prompt}]}]}
|
|
logger.info("Sende Anfrage an Gemini API...")
|
|
# logger.debug(f"Rohe Gemini API-Anfrage (JSON): {json.dumps(payload, indent=2)}")
|
|
try:
|
|
response = requests.post(GEMINI_API_URL, json=payload, headers={'Content-Type': 'application/json'})
|
|
response.raise_for_status()
|
|
res_json = response.json()
|
|
logger.info(f"Gemini API-Antwort erhalten (Status: {response.status_code}).")
|
|
|
|
text = res_json['candidates'][0]['content']['parts'][0]['text']
|
|
result = _extract_json_from_text(text)
|
|
|
|
if not result:
|
|
raise ValueError("Konnte kein valides JSON extrahieren")
|
|
|
|
return result
|
|
|
|
except Exception as e:
|
|
logger.error(f"Strategy generation failed: {e}")
|
|
# Return fallback to avoid frontend crash
|
|
return {
|
|
"summaryOfOffer": "Error generating strategy. Please check logs.",
|
|
"idealCustomerProfile": "Error generating ICP. Please check logs.",
|
|
"searchStrategyICP": "Error generating Search Strategy ICP. Please check logs.",
|
|
"digitalSignals": "Error generating Digital Signals. Please check logs.",
|
|
"targetPages": "Error generating Target Pages. Please check logs.",
|
|
"signals": []
|
|
}
|
|
|
|
def identify_competitors(reference_url, target_market, industries, summary_of_offer=None):
|
|
logger.info(f"Identifying competitors for {reference_url}")
|
|
api_key = load_gemini_api_key()
|
|
# Switch to stable 2.5-pro model
|
|
GEMINI_API_URL = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-pro:generateContent?key={api_key}"
|
|
|
|
prompt = f"""
|
|
You are a B2B Market Analyst. Find 3-5 direct competitors or highly similar companies (lookalikes) for the company at `{reference_url}`.
|
|
|
|
--- CONTEXT ---
|
|
- Offer: {summary_of_offer}
|
|
- Target Market: {target_market}
|
|
- Relevant Industries: {', '.join(industries)}
|
|
|
|
--- TASK ---
|
|
Identify competitors and categorize them into three groups:
|
|
1. 'localCompetitors': Competitors in the same immediate region/city.
|
|
2. 'nationalCompetitors': Competitors operating across the same country.
|
|
3. 'internationalCompetitors': Global players.
|
|
|
|
For EACH competitor, you MUST provide:
|
|
- `id`: A unique, URL-friendly identifier (e.g., "competitor-name-gmbh").
|
|
- `name`: The official, full name of the company.
|
|
- `description`: A concise explanation of why they are a competitor.
|
|
|
|
--- OUTPUT FORMAT ---
|
|
Return ONLY a valid JSON object with the following structure:
|
|
{{
|
|
"localCompetitors": [ {{ "id": "...", "name": "...", "description": "..." }} ],
|
|
"nationalCompetitors": [ ... ],
|
|
"internationalCompetitors": [ ... ]
|
|
}}
|
|
"""
|
|
|
|
payload = {"contents": [{"parts": [{"text": prompt}]}]}
|
|
logger.info("Sende Anfrage an Gemini API...")
|
|
# logger.debug(f"Rohe Gemini API-Anfrage (JSON): {json.dumps(payload, indent=2)}")
|
|
try:
|
|
response = requests.post(GEMINI_API_URL, json=payload, headers={'Content-Type': 'application/json'})
|
|
response.raise_for_status()
|
|
res_json = response.json()
|
|
logger.info(f"Gemini API-Antwort erhalten (Status: {response.status_code}).")
|
|
|
|
text = res_json['candidates'][0]['content']['parts'][0]['text']
|
|
result = _extract_json_from_text(text)
|
|
|
|
if not result:
|
|
raise ValueError("Konnte kein valides JSON extrahieren")
|
|
|
|
return result
|
|
|
|
except Exception as e:
|
|
logger.error(f"Competitor identification failed: {e}")
|
|
return {"localCompetitors": [], "nationalCompetitors": [], "internationalCompetitors": []}
|
|
|
|
def analyze_company(company_name, strategy, target_market):
|
|
logger.info(f"--- STARTING DEEP TECH AUDIT FOR: {company_name} ---")
|
|
api_key = load_gemini_api_key()
|
|
# Switch to stable 2.5-pro model
|
|
GEMINI_API_URL = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-pro:generateContent?key={api_key}"
|
|
|
|
# 1. Website Finding (SerpAPI fallback to Gemini)
|
|
url = None
|
|
website_search_results = serp_search(f"{company_name} offizielle Website")
|
|
if website_search_results:
|
|
url = website_search_results[0].get("link")
|
|
logger.info(f"Website via SerpAPI gefunden: {url}")
|
|
|
|
if not url:
|
|
# Fallback: Frage Gemini (Low Confidence)
|
|
logger.info("Keine URL via SerpAPI, frage Gemini...")
|
|
prompt_url = f"What is the official homepage URL for the company '{company_name}' in the market '{target_market}'? Respond with ONLY the single, complete URL and nothing else."
|
|
payload_url = {"contents": [{"parts": [{"text": prompt_url}]}]}
|
|
logger.info("Sende Anfrage an Gemini API (URL Fallback)...")
|
|
# logger.debug(f"Rohe Gemini API-Anfrage (JSON): {json.dumps(payload_url, indent=2)}")
|
|
try:
|
|
res = requests.post(GEMINI_API_URL, json=payload_url, headers={'Content-Type': 'application/json'}, timeout=15)
|
|
res.raise_for_status()
|
|
res_json = res.json()
|
|
logger.info(f"Gemini API-Antwort erhalten (Status: {res.status_code}).")
|
|
|
|
candidate = res_json.get('candidates', [{}])[0]
|
|
content = candidate.get('content', {}).get('parts', [{}])[0]
|
|
text_response = content.get('text', '').strip()
|
|
|
|
url_match = re.search(r'(https?://[^\s"]+)', text_response)
|
|
if url_match:
|
|
url = url_match.group(1)
|
|
logger.info(f"Gemini Fallback hat URL gefunden: {url}")
|
|
else:
|
|
logger.warning(f"Keine gültige URL in Gemini-Antwort gefunden: '{text_response}'")
|
|
|
|
except Exception as e:
|
|
logger.error(f"Gemini URL Fallback failed: {e}")
|
|
pass
|
|
|
|
if not url or not url.startswith("http"):
|
|
return {"error": f"Could not find website for {company_name}"}
|
|
|
|
# 2. Homepage Scraping
|
|
homepage_text = get_website_text(url)
|
|
if not homepage_text:
|
|
return {"error": f"Could not scrape website {url}"}
|
|
|
|
homepage_text = re.sub(r'[^\x20-\x7E\n\r\t]', '', homepage_text)
|
|
|
|
# --- ENHANCED: EXTERNAL TECHNOGRAPHIC INTELLIGENCE ---
|
|
# Suche aktiv nach Wettbewerbern, nicht nur auf der Firmenwebsite.
|
|
tech_evidence = []
|
|
|
|
# Liste bekannter Wettbewerber / Incumbents
|
|
known_incumbents = [
|
|
"SAP Ariba", "Jaggaer", "Coupa", "SynerTrade", "Ivalua",
|
|
"ServiceNow", "Salesforce", "Oracle SCM", "Zycus", "GEP",
|
|
"SupplyOn", "EcoVadis", "IntegrityNext"
|
|
]
|
|
|
|
# Suche 1: Direkte Verbindung zu Software-Anbietern (Case Studies, News, etc.)
|
|
# Wir bauen eine Query mit OR, um API-Calls zu sparen.
|
|
# Splitte in 2 Gruppen, um Query-Länge im Rahmen zu halten
|
|
half = len(known_incumbents) // 2
|
|
group1 = " OR ".join([f'"{inc}"' for inc in known_incumbents[:half]])
|
|
group2 = " OR ".join([f'"{inc}"' for inc in known_incumbents[half:]])
|
|
|
|
tech_queries = [
|
|
f'"{company_name}" ({group1})',
|
|
f'"{company_name}" ({group2})',
|
|
f'"{company_name}" "supplier portal" login' # Suche nach dem Portal selbst
|
|
]
|
|
|
|
logger.info(f"Starte erweiterte Tech-Stack-Suche für {company_name}...")
|
|
for q in tech_queries:
|
|
logger.info(f"Tech Search: {q}")
|
|
results = serp_search(q, num_results=4) # Etwas mehr Ergebnisse
|
|
if results:
|
|
for r in results:
|
|
tech_evidence.append(f"- Found: {r['title']}\n Snippet: {r['snippet']}\n Link: {r['link']}")
|
|
|
|
tech_evidence_text = "\n".join(tech_evidence)
|
|
# --- END ENHANCED TECH SEARCH ---
|
|
|
|
# 3. Targeted Signal Search (The "Hunter" Phase) - Basierend auf Strategy
|
|
signal_evidence = []
|
|
|
|
# Firmographics Search
|
|
firmographics_results = serp_search(f"{company_name} Umsatz Mitarbeiterzahl 2023")
|
|
firmographics_context = "\n".join([f"- {r['snippet']} ({r['link']})" for r in firmographics_results])
|
|
|
|
# Signal Searches (Original Strategy)
|
|
signals = strategy.get('signals', [])
|
|
for signal in signals:
|
|
# Überspringe Signale, die wir schon durch die Tech-Suche massiv abgedeckt haben,
|
|
# es sei denn, sie sind sehr spezifisch.
|
|
if "incumbent" in signal['id'].lower() or "tech" in signal['id'].lower():
|
|
logger.info(f"Skipping generic signal search '{signal['name']}' in favor of Enhanced Tech Search.")
|
|
continue
|
|
|
|
proof_strategy = signal.get('proofStrategy', {})
|
|
query_template = proof_strategy.get('searchQueryTemplate')
|
|
|
|
search_context = ""
|
|
if query_template:
|
|
try:
|
|
domain = url.split("//")[-1].split("/")[0].replace("www.", "")
|
|
except:
|
|
domain = ""
|
|
|
|
query = query_template.replace("{{COMPANY}}", company_name).replace("{COMPANY}", company_name)
|
|
query = query.replace("{{domain}}", domain).replace("{domain}", domain)
|
|
|
|
logger.info(f"Signal Search '{signal['name']}': {query}")
|
|
results = serp_search(query, num_results=3)
|
|
if results:
|
|
search_context = "\n".join([f" * Snippet: {r['snippet']}\n Source: {r['link']}" for r in results])
|
|
|
|
if search_context:
|
|
signal_evidence.append(f"SIGNAL '{signal['name']}':\n{search_context}")
|
|
|
|
# 4. Final Analysis & Synthesis (The "Judge" Phase)
|
|
evidence_text = "\n\n".join(signal_evidence)
|
|
|
|
prompt = f"""
|
|
You are a Strategic B2B Sales Consultant.
|
|
Analyze the company '{company_name}' ({url}) to create a "best-of-breed" sales pitch strategy.
|
|
|
|
--- STRATEGY (What we are looking for) ---
|
|
{json.dumps(signals, indent=2)}
|
|
|
|
--- EVIDENCE 1: EXTERNAL TECH-STACK INTELLIGENCE (CRITICAL) ---
|
|
Look closely here for mentions of competitors like SAP Ariba, Jaggaer, SynerTrade, Coupa, etc.
|
|
{tech_evidence_text}
|
|
|
|
--- EVIDENCE 2: HOMEPAGE CONTENT ---
|
|
{homepage_text[:8000]}
|
|
|
|
--- EVIDENCE 3: FIRMOGRAPHICS SEARCH ---
|
|
{firmographics_context}
|
|
|
|
--- EVIDENCE 4: TARGETED SIGNAL SEARCH RESULTS ---
|
|
{evidence_text}
|
|
----------------------------------
|
|
|
|
TASK:
|
|
1. **Firmographics**: Estimate Revenue and Employees.
|
|
2. **Technographic Audit**: Look for specific competitor software or legacy systems mentioned in EVIDENCE 1 (e.g., "Partner of SynerTrade", "Login to Jaggaer Portal").
|
|
3. **Status**:
|
|
- Set to "Nutzt Wettbewerber" if ANY competitor technology is found (Ariba, Jaggaer, SynerTrade, Coupa, etc.).
|
|
- Set to "Greenfield" ONLY if absolutely no competitor tech is found.
|
|
- Set to "Bestandskunde" if they already use our solution.
|
|
4. **Evaluate Signals**: For each signal, provide a "value" (Yes/No/Partial) and "proof".
|
|
5. **Recommendation (Pitch Strategy)**:
|
|
- DO NOT write a generic verdict.
|
|
- If they use a competitor (e.g., Ariba), explain how to position against it (e.g., "Pitch as a specialized add-on for logistics, filling Ariba's gaps").
|
|
- If Greenfield, explain the entry point.
|
|
- **Tone**: Strategic, insider-knowledge, specific.
|
|
|
|
STRICTLY output only JSON:
|
|
{{
|
|
"companyName": "{company_name}",
|
|
"status": "...",
|
|
"revenue": "...",
|
|
"employees": "...",
|
|
"tier": "Tier 1/2/3",
|
|
"dynamicAnalysis": {{
|
|
"sig_id_from_strategy": {{ "value": "...", "proof": "..." }}
|
|
}},
|
|
"recommendation": "..."
|
|
}}
|
|
"""
|
|
|
|
payload = {
|
|
"contents": [{"parts": [{"text": prompt}]}],
|
|
"generationConfig": {"response_mime_type": "application/json"}
|
|
}
|
|
|
|
try:
|
|
logger.info("Sende Audit-Anfrage an Gemini API...")
|
|
# logger.debug(f"Rohe Gemini API-Anfrage (JSON): {json.dumps(payload, indent=2)}")
|
|
response = requests.post(GEMINI_API_URL, json=payload, headers={'Content-Type': 'application/json'})
|
|
response.raise_for_status()
|
|
response_data = response.json()
|
|
logger.info(f"Gemini API-Antwort erhalten (Status: {response.status_code}).")
|
|
|
|
text = response_data['candidates'][0]['content']['parts'][0]['text']
|
|
result = _extract_json_from_text(text)
|
|
|
|
if not result:
|
|
raise ValueError("Konnte kein valides JSON extrahieren")
|
|
|
|
result['dataSource'] = "Digital Trace Audit (Deep Dive)"
|
|
logger.info(f"Audit für {company_name} erfolgreich abgeschlossen.")
|
|
return result
|
|
except Exception as e:
|
|
logger.error(f"Audit failed for {company_name}: {e}")
|
|
return {
|
|
"companyName": company_name,
|
|
"status": "Unklar / Manuelle Prüfung",
|
|
"revenue": "Error",
|
|
"employees": "Error",
|
|
"tier": "Tier 3",
|
|
"dynamicAnalysis": {},
|
|
"recommendation": f"Audit failed due to API Error: {str(e)}",
|
|
"dataSource": "Error"
|
|
}
|
|
|
|
def generate_outreach_campaign(company_data_json, knowledge_base_content, reference_url):
|
|
"""
|
|
Erstellt personalisierte E-Mail-Kampagnen basierend auf Audit-Daten und einer strukturierten Wissensdatenbank.
|
|
Generiert spezifische Ansprachen für verschiedene Rollen (Personas).
|
|
"""
|
|
company_name = company_data_json.get('companyName', 'Unknown')
|
|
logger.info(f"--- STARTING ROLE-BASED OUTREACH GENERATION FOR: {company_name} ---")
|
|
|
|
api_key = load_gemini_api_key()
|
|
# Switch to stable 2.5-pro model
|
|
GEMINI_API_URL = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-pro:generateContent?key={api_key}"
|
|
|
|
prompt = f"""
|
|
You are a Strategic Key Account Manager and deeply technical Industry Insider.
|
|
Your goal is to write highly personalized, **operationally specific** outreach emails to the company '{company_name}'.
|
|
|
|
--- INPUT 1: YOUR IDENTITY & STRATEGY (The Sender) ---
|
|
The following Markdown contains your company's identity, products, and strategy.
|
|
You act as the sales representative for the company described here:
|
|
{knowledge_base_content}
|
|
|
|
--- INPUT 2: THE TARGET COMPANY (Audit Facts) ---
|
|
{json.dumps(company_data_json, indent=2)}
|
|
|
|
--- INPUT 3: THE REFERENCE CLIENT (Social Proof) ---
|
|
Reference Client URL: {reference_url}
|
|
|
|
CRITICAL: This 'Reference Client' is an existing happy customer of ours. They are the "Seed Company" used to find the Target Company (Lookalike).
|
|
You MUST mention this Reference Client by name (derive it from the URL, e.g., 'schindler.com' -> 'Schindler') to establish trust.
|
|
|
|
--- TASK ---
|
|
1. **Analyze**: Match the Target Company (Input 2) to the most relevant 'Zielbranche/Segment' from the Knowledge Base (Input 1).
|
|
2. **Select Roles**: Identify the top 2 most distinct and relevant 'Rollen' (Personas) from the Knowledge Base for this specific company situation.
|
|
- *Example:* If the audit says they use a competitor (risk of lock-in), select a role like "Strategic Purchaser" or "Head of R&D" who cares about "Second Source".
|
|
- *Example:* If they have quality issues or complex logistics, pick "Quality Manager" or "Logistics Head".
|
|
3. **Draft Campaigns**: For EACH of the 2 selected roles, write a 3-step email sequence.
|
|
|
|
--- TONE & STYLE GUIDELINES (CRITICAL) ---
|
|
- **Perspective:** Operational Expert & Insider. NOT generic marketing.
|
|
- **Be Gritty & Specific:** Do NOT use fluff like "optimize efficiency" or "streamline processes" without context.
|
|
- Use **hard, operational keywords** from the Knowledge Base (e.g., "ASNs", "VMI", "8D-Reports", "Maverick Buying", "Bandstillstand", "Sonderfahrten", "PPAP").
|
|
- Show you understand their daily pain.
|
|
- **Narrative Arc:**
|
|
1. "I noticed [Fact from Audit/Tech Stack]..." (e.g., "You rely on PDF orders via Jaggaer...")
|
|
2. "In [Industry], this often leads to [Operational Pain]..." (e.g., "missing ASNs causing delays at the hub.")
|
|
3. "We helped [Reference Client Name] solve exactly this by [Specific Solution]..."
|
|
4. "Let's discuss how to get [Operational Gain] without replacing your ERP."
|
|
- **Mandatory Social Proof:** You MUST mention the Reference Client Name (from Input 3) in the email body or footer.
|
|
- **Language:** German (as the inputs are German).
|
|
|
|
--- OUTPUT FORMAT (Strictly JSON) ---
|
|
Returns a list of campaigns.
|
|
[
|
|
{{
|
|
"target_role": "Name of the Role (e.g. Leiter F&E)",
|
|
"rationale": "Why this role? (e.g. Because the audit found dependency on Competitor X...)",
|
|
"emails": [
|
|
{{
|
|
"subject": "Specific Subject Line",
|
|
"body": "Email Body..."
|
|
}},
|
|
{{
|
|
"subject": "Re: Subject",
|
|
"body": "Follow-up Body..."
|
|
}},
|
|
{{
|
|
"subject": "Final Check",
|
|
"body": "Final Body..."
|
|
}}
|
|
]
|
|
}},
|
|
... (Second Role)
|
|
]
|
|
"""
|
|
|
|
payload = {
|
|
"contents": [{"parts": [{"text": prompt}]}],
|
|
"generationConfig": {"response_mime_type": "application/json"}
|
|
}
|
|
|
|
try:
|
|
logger.info("Sende Campaign-Anfrage an Gemini API...")
|
|
# logger.debug(f"Rohe Gemini API-Anfrage (JSON): {json.dumps(payload, indent=2)}")
|
|
response = requests.post(GEMINI_API_URL, json=payload, headers={'Content-Type': 'application/json'})
|
|
response.raise_for_status()
|
|
response_data = response.json()
|
|
logger.info(f"Gemini API-Antwort erhalten (Status: {response.status_code}).")
|
|
# logger.debug(f"Rohe API-Antwort (JSON): {json.dumps(response_data, indent=2)}")
|
|
|
|
text = response_data['candidates'][0]['content']['parts'][0]['text']
|
|
result = _extract_json_from_text(text)
|
|
|
|
if not result:
|
|
raise ValueError("Konnte kein valides JSON extrahieren")
|
|
|
|
return result
|
|
except Exception as e:
|
|
logger.error(f"Campaign generation failed for {company_name}: {e}")
|
|
return [{"error": str(e)}]
|
|
|
|
def main():
|
|
parser = argparse.ArgumentParser()
|
|
parser.add_argument("--mode", required=True)
|
|
parser.add_argument("--reference_url")
|
|
parser.add_argument("--context_file")
|
|
parser.add_argument("--target_market")
|
|
parser.add_argument("--company_name")
|
|
parser.add_argument("--strategy_json")
|
|
parser.add_argument("--summary_of_offer")
|
|
parser.add_argument("--company_data_file") # For generate_outreach
|
|
args = parser.parse_args()
|
|
|
|
if args.mode == "generate_strategy":
|
|
with open(args.context_file, "r") as f: context = f.read()
|
|
print(json.dumps(generate_search_strategy(args.reference_url, context)))
|
|
elif args.mode == "identify_competitors":
|
|
industries = []
|
|
if args.context_file:
|
|
with open(args.context_file, "r") as f: context = f.read()
|
|
industries = _extract_target_industries_from_context(context)
|
|
print(json.dumps(identify_competitors(args.reference_url, args.target_market, industries, args.summary_of_offer)))
|
|
elif args.mode == "analyze_company":
|
|
strategy = json.loads(args.strategy_json)
|
|
print(json.dumps(analyze_company(args.company_name, strategy, args.target_market)))
|
|
elif args.mode == "generate_outreach":
|
|
with open(args.company_data_file, "r") as f: company_data = json.load(f)
|
|
with open(args.context_file, "r") as f: knowledge_base = f.read()
|
|
print(json.dumps(generate_outreach_campaign(company_data, knowledge_base, args.reference_url)))
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|