Merge pull request #231 from ConnectAI-E/freeziyou-022201

支持飞书传图GitHub

server/routes/team.py

@@ -175,29 +175,29 @@ def install_im_application_to_team_by_get_method(team_id, platform):
             )
             app.logger.info("result %r", result)
             events = [
-                "20",
-                "im.message.message_read_v1",
-                "im.message.reaction.created_v1",
-                "im.message.reaction.deleted_v1",
-                "im.message.recalled_v1",
-                "im.message.receive_v1",
+                "20",  # 用户和机器人的会话首次被创建
+                "im.message.message_read_v1",  # 消息已读
+                "im.message.reaction.created_v1",  # 新增消息表情回复
+                "im.message.reaction.deleted_v1",  # 删除消息表情回复
+                "im.message.recalled_v1",  # 撤回消息
+                "im.message.receive_v1",  # 接收消息
             ]
             scope_ids = [
-                "8002",
-                "100032",
-                "6081",
-                "14",
-                "1",
-                "21001",
-                "20001",
-                "20011",
-                "3001",
-                "20012",
-                "20010",
-                "3000",
-                "20008",
-                "1000",
-                "20009",
+                "8002",  # 获取应用信息
+                "100032",  # 获取通讯录基本信息
+                "6081",  # 以应用身份读取通讯录
+                "14",  # 获取用户基本信息
+                "1",  # 获取用户邮箱信息
+                "21001",  # 获取与更新群组信息
+                "20001",  # 获取与发送单聊、群组消息
+                "20011",  # 获取用户在群组中 @ 机器人的消息
+                "3001",  # 接收群聊中 @ 机器人消息事件
+                "20012",  # 获取群组中所有消息
+                "20010",  # 获取用户发给机器人的单聊消息
+                "3000",  # 读取用户发给机器人的单聊消息
+                "20008",  # 获取单聊、群组消息
+                "1000",  # 以应用的身份发消息
+                "20009",  # 获取上传图片或文件资源
             ]
             hook_url = f"{os.environ.get('DOMAIN')}/api/feishu/hook/{app_id}"
             return redirect(
@@ -260,9 +260,9 @@ def get_task_result_by_id(team_id, task_id):
             "data": {
                 "task_id": task.id,
                 "status": task.status,
-                "result": task.result
-                if isinstance(task.result, list)
-                else str(task.result),
+                "result": (
+                    task.result if isinstance(task.result, list) else str(task.result)
+                ),
             },
         }
     )

server/routes/user.py

@@ -1,8 +1,14 @@
 from app import app
-from flask import Blueprint, jsonify, request, session
-from model.team import get_team_list_by_user_id, is_team_admin
+from flask import Blueprint, Response, abort, jsonify, request, session
+from model.team import (
+    get_application_info_by_team_id,
+    get_team_list_by_user_id,
+    is_team_admin,
+)
 from model.user import get_user_by_id
+from tasks.lark.base import get_bot_by_application_id, get_repo_by_repo_id
 from utils.auth import authenticated
+from utils.utils import download_file
 
 bp = Blueprint("user", __name__, url_prefix="/api")
 
@@ -56,4 +62,38 @@ def set_account():
     return jsonify({"code": 0, "msg": "success"})
 
 
+@bp.route("/<team_id>/<repo_id>/<message_id>/image/<img_key>", methods=["GET"])
+def get_image(team_id, message_id, repo_id, img_key):
+    """
+    1. 用 img_key 请求飞书接口下载 image
+    2. 判断请求来源，如果是 GitHub 调用，则直接返回 image
+    3. 用户调用 校验权限
+    """
+
+    def download_and_respond():
+        _, im_application = get_application_info_by_team_id(team_id)
+        bot, _ = get_bot_by_application_id(im_application.id)
+        image_content = download_file(img_key, message_id, bot, "image")
+        return Response(image_content, mimetype="image/png")
+
+    # GitHub调用
+    user_agent = request.headers.get("User-Agent")
+    if user_agent and user_agent.startswith("github-camo"):
+        return download_and_respond()
+
+    # TODO 用户调用(弱需求, 通常来讲此接口不会被暴露), 需要进一步校验权限
+    referer = request.headers.get("Referer")
+    if not referer:
+        # 公开仓库不校验
+        repo = get_repo_by_repo_id(repo_id)
+        is_private = repo.extra.get("private", False)
+        app.logger.debug(f"is_private: {is_private}")
+
+        # 私有仓库校验，先登录
+        if is_private:
+            return abort(403)
+
+    return download_and_respond()
+
+
 app.register_blueprint(bp)

server/tasks/lark/base.py

@@ -24,6 +24,11 @@ def get_chat_group_by_chat_id(chat_id):
 
 
 def get_repo_name_by_repo_id(repo_id):
+    repo = get_repo_by_repo_id(repo_id)
+    return repo.name
+
+
+def get_repo_by_repo_id(repo_id):
     repo = (
         db.session.query(Repo)
         .filter(
@@ -32,7 +37,7 @@ def get_repo_name_by_repo_id(repo_id):
         )
         .first()
     )
-    return repo.name
+    return repo
 
 
 def get_bot_by_application_id(app_id):

server/tasks/lark/chat.py

@@ -1,5 +1,7 @@
 import json
 import logging
+import os
+import re
 from urllib.parse import urlparse
 
 from celery_app import app, celery
@@ -313,13 +315,8 @@ def create_issue(
     )
     assignees = [code_users[openid][1] for openid in users if openid in code_users]
 
-    # 判断 content 中是否有 at
-    if "mentions" in data["event"]["message"]:
-        # 替换 content 中的 im_name 为 code_name
-        body = replace_im_name_to_github_name(
-            app_id, message_id, {"text": body}, data, team, *args, **kwargs
-        )
-        body = body.replace("\n", "\r\n")
+    # 处理 body
+    body = process_desc(app_id, message_id, repo.id, body, data, team, *args, **kwargs)
 
     response = github_app.create_issue(
         team.name, repo.name, title, body, assignees, labels
@@ -331,6 +328,44 @@ def create_issue(
     return response
 
 
+def process_desc(app_id, message_id, repo_id, desc, data, team, *args, **kwargs):
+    """
+    处理发给 github 的 desc, 转换@、处理图片、换行
+    """
+    # 1. 判断 body 中是否有 at
+    if "mentions" in data["event"]["message"]:
+        # 替换 body 中的 im_name 为 code_name
+        desc = replace_im_name_to_github_name(
+            app_id, message_id, {"text": desc}, data, team, *args, **kwargs
+        )
+
+    # 2. 处理 body 中的图片
+    desc = replace_images_keys_with_url(desc, team.id, message_id, repo_id)
+
+    # github 只支持 \r\n
+    return desc.replace("\n", "\r\n")
+
+
+def replace_images_keys_with_url(text, team_id, message_id, repo_id):
+    """
+    replace image_key with image URL.
+    ![](image_key) -> ![](gitmaya.com/api/<team_id>/<repo_id>/<message_id>/image/<image_key>)
+    Args:
+        text (str): original text
+
+    Returns:
+        str: replaced text
+    """
+    host = os.environ.get("DOMAIN")
+    replaced_text = re.sub(
+        r"!\[.*?\]\((.*?)\)",
+        lambda match: f"![]({host}/api/{team_id}/{repo_id}/{message_id}/image/{match.group(1)})",
+        text,
+    )
+
+    return replaced_text
+
+
 @celery.task()
 def sync_issue(
     issue_id, issue_link, app_id, message_id, content, data, *args, **kwargs

server/tasks/lark/issue.py

@@ -21,7 +21,7 @@
 from utils.lark.issue_manual_help import IssueManualHelp, IssueView
 from utils.lark.issue_tip_failed import IssueTipFailed
 from utils.lark.issue_tip_success import IssueTipSuccess
-from utils.utils import upload_image
+from utils.utils import process_image
 
 from .base import (
     get_bot_by_application_id,
@@ -122,12 +122,12 @@ def gen_issue_card_by_issue(bot, issue, repo_url, team, maunal=False):
             tags=tags,
         )
 
-    # 处理 description 中的图片
+    # 处理从 github 创建 Issue 时, description 中的图片
     description = replace_images_with_keys(
         issue.description if issue.description else "", bot
     )
 
-    # 处理 description 中的at
+    # 处理从 github 创建 Issue 时, description 中的 at
     description = replace_code_name_to_im_name(description)
 
     return IssueCard(
@@ -160,15 +160,15 @@ def replace_images_with_keys(text, bot):
     markdown_pattern = r"!\[.*?\]\((.*?)\)"
     replaced_text = re.sub(
         markdown_pattern,
-        lambda match: f"![]({upload_image(match.group(1), bot)})",
+        lambda match: f"![]({process_image(match.group(1), bot)})",
         text,
     )
 
     # Replace HTML image syntax
     html_pattern = r"<img.*?src=\"(.*?)\".*?>"
     replaced_text = re.sub(
         html_pattern,
-        lambda match: f"![]({upload_image(match.group(1), bot)})",
+        lambda match: f"![]({process_image(match.group(1), bot)})",
         replaced_text,
     )
 
@@ -601,12 +601,12 @@ def create_issue_comment(app_id, message_id, content, data, *args, **kwargs):
     )
     comment_text = content["text"]
 
-    # 判断 content 中是否有 at
-    if "mentions" in data["event"]["message"]:
-        # 替换 content 中的 im_name 为 code_name
-        comment_text = replace_im_name_to_github_name(
-            app_id, message_id, content, data, team, *args, **kwargs
-        )
+    from tasks.lark.chat import process_desc
+
+    # 处理 desc
+    comment_text = process_desc(
+        app_id, message_id, repo.id, comment_text, data, team, *args, **kwargs
+    )
 
     response = github_app.create_issue_comment(
         team.name, repo.name, issue.issue_number, comment_text

server/utils/utils.py

@@ -1,15 +1,23 @@
 import logging
+import os
 
 import httpx
 from utils.redis import stalecache
 
 
+def process_image(url, bot):
+    if not url or not url.startswith("http"):
+        return ""
+
+    if url.startswith(f"{os.environ.get('DOMAIN')}/api"):
+        return url.split("/")[-1]
+    return upload_image(url, bot)
+
+
 # 使用 stalecache 装饰器，以 url 作为缓存键
 @stalecache(expire=3600, stale=600)
 def upload_image(url, bot):
     logging.info("upload image: %s", url)
-    if not url or not url.startswith("http"):
-        return ""
     response = httpx.get(url, follow_redirects=True)
     if response.status_code == 200:
         # 函数返回值: iamg_key 存到缓存中
@@ -30,6 +38,18 @@ def upload_image_binary(img_bin, bot):
     return response["data"]["image_key"]
 
 
+@stalecache(expire=3600, stale=600)
+def download_file(file_key, message_id, bot, file_type="image"):
+    """
+    获取消息中的资源文件，包括音频，视频，图片和文件，暂不支持表情包资源下载。当前仅支持 100M 以内的资源文件的下载
+    """
+    # open-apis/im/v1/images/{img_key} 接口只能下载机器人自己上传的图片
+    url = f"{bot.host}/open-apis/im/v1/messages/{message_id}/resources/{file_key}?type={file_type}"
+
+    response = bot.get(url)
+    return response.content
+
+
 def query_one_page(query, page, size):
     offset = (page - 1) * int(size)
     return (