feat(content): implement Content Engine MVP (v1.0) with GTM integration

content-engine/content_orchestrator.py

@@ -0,0 +1,181 @@
+
+import argparse
+import base64
+import json
+import logging
+import sys
+import os
+from datetime import datetime
+import content_db_manager as db_manager
+
+# Ensure helper path is correct
+sys.path.append(os.path.dirname(os.path.abspath(__file__)))
+sys.path.append(os.path.join(os.path.dirname(os.path.abspath(__file__)), ".."))
+
+from helpers import call_gemini_flash, scrape_website_details
+from config import Config
+
+LOG_DIR = "Log_from_docker"
+if not os.path.exists(LOG_DIR):
+    os.makedirs(LOG_DIR)
+
+run_timestamp = datetime.now().strftime("%y-%m-%d_%H-%M-%S")
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+
+Config.load_api_keys()
+
+def get_copywriter_instruction(lang='de'):
+    return r"""
+    Du bist ein Senior Copywriter und SEO-Experte. Deine Spezialität ist der 'Challenger Sale'.
+    Du schreibst Texte, die fachlich tief fundiert, professionell und leicht aggressiv/fordernd sind.
+    
+    DEIN STIL:
+    - Keine Buzzwords ohne Substanz.
+    - Fokus auf den 'Cost of Inaction' (Was kostet es den Kunden, wenn er NICHT handelt?).
+    - Übersetzung von Technik in geschäftlichen Nutzen.
+    - SEO-Integration: Baue Keywords natürlich aber präsent ein.
+    """
+
+# --- MODES ---
+
+def list_gtm_projects(payload):
+    projects = db_manager.get_all_gtm_projects()
+    return {"projects": projects}
+
+def import_project(payload):
+    gtm_id = payload.get('gtmProjectId')
+    if not gtm_id:
+        return {"error": "Missing gtmProjectId"}
+    
+    result = db_manager.import_gtm_project(gtm_id)
+    if not result:
+        return {"error": "GTM Project not found or import failed"}
+    
+    return result
+
+def list_content_projects(payload):
+    projects = db_manager.get_all_content_projects()
+    return {"projects": projects}
+
+def load_project_details(payload):
+    project_id = payload.get('projectId')
+    project = db_manager.get_content_project(project_id)
+    if not project:
+        return {"error": "Project not found"}
+    
+    assets = db_manager.get_project_assets(project_id)
+    project['assets'] = assets
+    return project
+
+def seo_brainstorming(payload):
+    project_id = payload.get('projectId')
+    lang = payload.get('lang', 'de')
+    
+    project = db_manager.get_content_project(project_id)
+    if not project:
+        return {"error": "Project context not found"}
+    
+    gtm_data = project.get('gtm_data_snapshot', {})
+    
+    # GOLDEN RULE: Use Raw Quotes and .format()
+    prompt = r"""
+    Basierend auf folgendem GTM-Kontext (Strategie für ein technisches Produkt):
+    {gtm_context}
+    
+    AUFGABE:
+    Generiere eine strategische Liste von 15 SEO-Keywords. 
+    1. 5 Short-Tail Fokus-Keywords (z.B. Produktkategorie + 'kaufen/mieten').
+    2. 10 Long-Tail Keywords, die spezifische Pain Points oder Usecases adressieren (z.B. 'Kostenreduktion bei Sicherheitsrundgängen').
+    
+    Die Keywords müssen für Entscheider relevant sein (CFO, Head of Security, Operations Manager).
+    
+    Output NUR als JSON Liste von Strings.
+    """.format(gtm_context=json.dumps(gtm_data))
+    
+    response = call_gemini_flash(prompt, system_instruction=get_copywriter_instruction(lang), json_mode=True)
+    keywords = json.loads(response)
+    
+    db_manager.save_seo_strategy(project_id, {"seed_keywords": keywords})
+    return {"keywords": keywords}
+
+def generate_section(payload):
+    project_id = payload.get('projectId')
+    section_key = payload.get('sectionKey') # e.g., 'hero', 'problem', 'features'
+    manual_content = payload.get('manualContent')
+    lang = payload.get('lang', 'de')
+    keywords = payload.get('keywords', [])
+    
+    if manual_content:
+        # User is saving their manual edits
+        db_manager.save_content_asset(project_id, 'website_section', section_key, f"Section: {section_key}", manual_content, keywords)
+        return {"status": "saved", "sectionKey": section_key}
+
+    project = db_manager.get_content_project(project_id)
+    if not project:
+        return {"error": "Project context not found"}
+        
+    gtm_data = project.get('gtm_data_snapshot', {})
+    
+    # Context extraction
+    category = project.get('category')
+    
+    prompt = r"""
+    Erstelle den Website-Inhalt für die Sektion '{section}' eines Produkts in der Kategorie '{cat}'.
+    
+    STRATEGIE-KONTEXT:
+    {gtm_context}
+    
+    SEO-KEYWORDS ZU NUTZEN:
+    {kws}
+    
+    ANFORDERUNG:
+    - Schreibe im Stil eines Senior Copywriters (fachlich fundiert, Challenger Sale).
+    - Format: Markdown.
+    - Die Sektion muss den Nutzer zur nächsten Aktion (CTA) führen.
+    """.format(
+        section=section_key,
+        cat=category,
+        gtm_context=json.dumps(gtm_data),
+        kws=json.dumps(keywords)
+    )
+    
+    content = call_gemini_flash(prompt, system_instruction=get_copywriter_instruction(lang), json_mode=False)
+    
+    # Save as asset
+    db_manager.save_content_asset(project_id, 'website_section', section_key, f"Section: {section_key}", content, keywords)
+    
+    return {"content": content, "sectionKey": section_key}
+
+def main():
+    parser = argparse.ArgumentParser(description="Content Engine Orchestrator")
+    parser.add_argument("--mode", required=True)
+    parser.add_argument("--payload_file", help="Path to JSON payload")
+    
+    args = parser.parse_args()
+    
+    payload = {}
+    if args.payload_file:
+        with open(args.payload_file, 'r') as f:
+            payload = json.load(f)
+            
+    modes = {
+        "list_gtm_projects": list_gtm_projects,
+        "import_project": import_project,
+        "list_content_projects": list_content_projects,
+        "load_project": load_project_details,
+        "seo_brainstorming": seo_brainstorming,
+        "generate_section": generate_section,
+    }
+    
+    if args.mode in modes:
+        try:
+            result = modes[args.mode](payload)
+            print(json.dumps(result, ensure_ascii=False))
+        except Exception as e:
+            logging.error(f"Error in mode {args.mode}: {str(e)}")
+            print(json.dumps({"error": str(e)}))
+    else:
+        print(json.dumps({"error": f"Unknown mode: {args.mode}"}))
+
+if __name__ == "__main__":
+    main()