Gemini Web: Context verified. Verifying the Dynamic Prompt Update Okay, let's get this pumpkin harvested! 🎃 You want pip.nbup() to be smarter and use the correct sample prompt when syncing notebooks back to the assets/nbs/ directory. Right now, it's hardcoded to always use the FAQuilizer prompt. Here's an implementation plan suitable for handing over to the Gemini CLI: Implementation Plan: Dynamic Prompt Syncing in pip.nbup() Goal: Modify pipulate/core.py so that pip.nbup() selects the appropriate sample prompt based on the notebook being synced (FAQuilizer.ipynb vs. URLinspector.ipynb) before writing to assets/nbs/. -1. 📚 Define Both Sample Prompts in core.py Locate the existing SAMPLE_PROMPT_SOURCE list within the nbup method in pipulate/core.py. Rename it to SAMPLE_PROMPT_SOURCE_FAQ for clarity. Add a new list named SAMPLE_PROMPT_SOURCE_URLI containing the source lines for the URLinspector prompt. # Inside pipulate/core.py, within the nbup method# --- Define Sample Data for Scrubbing ---SAMPLE_PROMPT_SOURCE_FAQ = [ "**Your Role (AI Content Strategist):**\n", "\n", "You are an AI Content Strategist. \n", "Make 5 Frequently Asked Questions for each page.\n", "For each question, produce the following so it fits the data structure:\n", "\n", "1. priority: integer (1-5, 1 is highest)\n", "2. question: string (The generated question)\n", "3. target_intent: string (What is the user's goal in asking this?)\n", "4. justification: string (Why is this a valuable question to answer? e.g., sales, seasonal, etc.)"]# ADD THIS NEW PROMPT LISTSAMPLE_PROMPT_SOURCE_URLI = [ "**Your Role (SEO URL Consultant auditing individual URLs):**\n", "\n", 'A "keyword" is the unique short combination of words that are most likely to be searched-on by a user who might be interested in this page.\n', 'What is the five most likely keywords (short combination of words) that this page appears to be targeting?\n', '\n', '1. priority: integer (1-5, 1 is highest)\n', '2. question: string (The keyword)\n', '3. target_intent: string (What specifically is the user likely to have in mind when performing this search?)\n', '4. justification: string (Informational, Commercial, Navigational, Transactional?)']SAMPLE_URL_LIST_SOURCE = [ # ... (keep existing URL list source)] -2. 🗺️ Create a Prompt Mapping Dictionary Still within the nbup method, create a dictionary that maps the base name of the notebook (without .ipynb) to the corresponding prompt source list. # Inside pipulate/core.py, within the nbup method, after defining the prompt sources# Map notebook base names to their sample promptsPROMPT_MAP = { "FAQuilizer": SAMPLE_PROMPT_SOURCE_FAQ, "URLinspector": SAMPLE_PROMPT_SOURCE_URLI, # Add mappings for other notebooks with prompts here} -3. \<0xF0>\<0x9F>\<0xAA>\<0x9A> Modify the nbup Scrubbing Logic Find the loop within nbup that iterates through the notebook cells (for cell in nb.cells:). Inside the loop, before checking the cell tags, determine the base name of the current notebook being processed. Use this base name to look up the correct prompt_source_to_use from the PROMPT_MAP. Provide a default fallback if needed. Modify the if "prompt-input" in tags: block to use prompt_source_to_use instead of the hardcoded SAMPLE_PROMPT_SOURCE. # Inside pipulate/core.py, within the nbup method's loop # --- Scrub proprietary data --- # <<< ADD THIS >>> notebook_base_name = Path(notebook_filename).stem # Get "URLinspector" from "URLinspector.ipynb" # Use the map, fallback to FAQ if notebook name not found (or choose another default) prompt_source_to_use = PROMPT_MAP.get(notebook_base_name, SAMPLE_PROMPT_SOURCE_FAQ) # <<< END ADDITION >>> for cell in nb.cells: tags = cell.metadata.get("tags", []) if "prompt-input" in tags: # <<< CHANGE THIS LINE >>> # cell.source = SAMPLE_PROMPT_SOURCE # Old hardcoded version cell.source = prompt_source_to_use # New dynamic version # <<< END CHANGE >>> print(f" ✓ Scrubbed and replaced 'prompt-input' cell using prompt for '{notebook_base_name}'.") elif "url-list-input" in tags: cell.source = SAMPLE_URL_LIST_SOURCE # This remains unchanged print(" ✓ Scrubbed and replaced 'url-list-input' cell.") # ... (rest of the cell processing logic remains the same) ...