پاسخ نظرات سایت وردپرس با api جمینای و گوگل

pip install google-genai pandas openpyxl

import sys
import xml.etree.ElementTree as ET
from typing import List, Dict
import pandas as pd
from datetime import datetime
import time
import os
import json

from google import genai
from google.genai import types

try:
    client = genai.Client()
except Exception as e:
    print(
        "Error initializing Gemini client. Make sure GEMINI_API_KEY is set.",
        file=sys.stderr,
    )
    pass

# GEMINI_MODEL = "gemini-2.0-flash-lite"
# REQUESTS_PER_MINUTE = 29

# GEMINI_MODEL = "gemini-2.5-flash"
# REQUESTS_PER_MINUTE = 10

# GEMINI_MODEL = "gemini-2.0-flash-exp"
# REQUESTS_PER_MINUTE = 10

GEMINI_MODEL = "gemini-2.5-flash-lite"
REQUESTS_PER_MINUTE = 15

# GEMINI_MODEL = "gemini-2.0-flash"
# REQUESTS_PER_MINUTE = 15

# GEMINI_MODEL = "gemini-2.5-pro"
# REQUESTS_PER_MINUTE = 2

# GEMINI_MODEL = "learnlm-2.0-flash-experimental"
# REQUESTS_PER_MINUTE = 15

CommentData = Dict[str, str]
CHECKPOINT_FILE = "checkpoint.json"


def load_xml(filename: str) -> ET.Element:
    try:
        tree = ET.parse(filename)
        return tree.getroot()
    except FileNotFoundError:
        print(f"Error: File '{filename}' not found.", file=sys.stderr)
        sys.exit(1)
    except ET.ParseError:
        print(f"Error: Could not parse XML from '{filename}'.", file=sys.stderr)
        sys.exit(1)


def extract_comments(root: ET.Element) -> List[CommentData]:
    ns = {
        "wp": "http://wordpress.org/export/1.2/",
        "content": "http://purl.org/rss/1.0/modules/content/",
    }
    comments_to_approve: List[CommentData] = []

    for item in root.findall("channel/item"):
        post_title = (
            item.find("title").text if item.find("title") is not None else "No Title"
        )
        post_content_elem = item.find("content:encoded", ns)
        post_content = (
            post_content_elem.text if post_content_elem is not None else "No Content"
        )

        post_id_elem = item.find("wp:post_id", ns)
        post_id = post_id_elem.text if post_id_elem is not None else "Anonymous"

        for comment in item.findall("wp:comment", ns):
            approved_elem = comment.find("wp:comment_approved", ns)
            approved = approved_elem.text if approved_elem is not None else "1"

            if approved == "0":
                comment_author_elem = comment.find("wp:comment_author", ns)
                comment_author = (
                    comment_author_elem.text
                    if comment_author_elem is not None
                    else "Anonymous"
                )

                comment_content_elem = comment.find("wp:comment_content", ns)
                comment_content = (
                    comment_content_elem.text
                    if comment_content_elem is not None
                    else "No Comment Content"
                )

                # Extract comment date
                comment_date_elem = comment.find("wp:comment_date", ns)
                comment_date_text = (
                    comment_date_elem.text if comment_date_elem is not None else ""
                )
                try:
                    comment_date = datetime.strptime(
                        comment_date_text, "%Y-%m-%d %H:%M:%S"
                    )
                except Exception:
                    comment_date = datetime.min  # If date is invalid

                comments_to_approve.append(
                    {
                        "post_id": post_id,
                        "post_title": post_title,
                        "post_content": post_content,
                        "comment_author": comment_author,
                        "comment_content": comment_content,
                        "comment_date": comment_date,
                    }
                )
    return comments_to_approve


def generate_reply(comment: CommentData) -> str:
    prompt = f"""
شما یک پست فارسی دارید:
عنوان پست:
{comment['post_title']}

محتوای کامل پست:
{comment['post_content']}

یک کامنت فارسی برای این پست دریافت شده:
نام نویسنده کامنت: {comment['comment_author']}
متن کامنت:
{comment['comment_content']}

لطفاً یک پاسخ **کوتاه، مودبانه و مرتبط** در حد یک یا دو جمله به این کامنت تولید کنید. پاسخ باید به زبان فارسی باشد. اسم نویسنده کامنت رو توی جوابت ننویس.
"""
    try:
        response = client.models.generate_content(
            model=GEMINI_MODEL,
            contents=prompt,
            config=types.GenerateContentConfig(
                max_output_tokens=150,
                temperature=0.4,
            ),
        )
        if response.text and response.text.strip():
            return response.text.strip()
        elif response.candidates:
            candidate = response.candidates[0]
            finish_reason = candidate.finish_reason.name
            if finish_reason == "SAFETY":
                return "❌ Response blocked due to safety filter."
            elif finish_reason == "MAX_TOKENS":
                return "❌ Response incomplete due to token limit."
            else:
                return f"❌ Error: Finish reason: {finish_reason}"
        else:
            return "❌ No response generated."
    except Exception as e:
        return f"❌ API Error: {e}"


def load_checkpoint() -> List[Dict]:
    if os.path.exists(CHECKPOINT_FILE):
        with open(CHECKPOINT_FILE, "r", encoding="utf-8") as f:
            return json.load(f)
    return []


def save_checkpoint(data: List[Dict]):
    with open(CHECKPOINT_FILE, "w", encoding="utf-8") as f:
        json.dump(data, f, default=str, ensure_ascii=False)


def main():
    if len(sys.argv) != 2:
        print('Usage: python answer.py "WordPress.xml"')
        sys.exit(1)

    filename = sys.argv[1]
    root = load_xml(filename)
    comments = extract_comments(root)

    if not comments:
        print("No pending comments found (wp:comment_approved = '0').")
        return

    comments.sort(key=lambda x: x["comment_date"])

    completed_replies = load_checkpoint()
    completed_post_ids = {r["post_id"] + r["comment_author"] for r in completed_replies}

    pending_comments = [
        c
        for c in comments
        if (c["post_id"] + c["comment_author"]) not in completed_post_ids
    ]

    print(f"Found {len(pending_comments)} pending comments. Generating replies...")
    print("=" * 60)

    for idx, c in enumerate(pending_comments, start=1):
        reply = generate_reply(c)
        completed_replies.append(
            {
                "post_id": c["post_id"],
                "post_title": c["post_title"],
                "comment_author": c["comment_author"],
                "comment_content": c["comment_content"],
                "comment_date": str(c["comment_date"]),
                "reply": reply,
            }
        )
        print(f"💬 Reply to {c['comment_author']}: {reply}")

        save_checkpoint(completed_replies)

        if idx % REQUESTS_PER_MINUTE == 0:
            print("⏳ Waiting to respect the 1-minute rate limit...")
            time.sleep(60)

    df = pd.DataFrame(completed_replies)
    df.to_excel("comment_replies.xlsx", index=False)
    print("✅ All replies saved to 'comment_replies.xlsx'.")


if __name__ == "__main__":
    main()

export GEMINI_API_KEY="APIKEYTEST"

python answer.py WordPress.xml