WangShen007
54b949b392
The upstream project name is gone from code, assets, docs and build
output: brand strings, domains (replypier.site and its app/api/docs/help
subdomains), env prefixes, image and network names, the widget's public
JS API (window.ReplyPier, replypierId/BaseUrl/Config/Token), CSS tokens
(--rp-*), file and directory names, and the WordPress plugin package.
- widget loader is now webclient/replypier.min.js; both bundles rebuilt
from source (backend/assets/widget.{js,css}, frontend/public/webclient)
- NOTICE, llms.txt, CONTRIBUTING, HELP_CENTER_INFRA and config.dev.js
rewritten rather than string-swapped; the demo config no longer ships
the previous vendor's Firebase/Razorpay/Google keys
- new migration replypier_ai_model_001 adds REPLYPIER to the managed
AIModelType and rewrites pre-rename rows
- local compose project, database, role and volume renamed; LOCAL_DEPLOY
documents carrying the old local data over
89 lines
3.3 KiB
Python
89 lines
3.3 KiB
Python
"""
|
|
Copyright 2024-2026 ReplyPier
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
|
|
Keep crawlers out of the API origin.
|
|
|
|
The API host had no robots.txt at all, so search engines were free to crawl
|
|
(and then report as 404s) endpoints that were never pages: /api/v1/... routes
|
|
and the YOUR_WIDGET_ID placeholder our docs use in examples.
|
|
|
|
Two halves, because they do different jobs. robots.txt stops the fetch;
|
|
X-Robots-Tag stops a URL discovered elsewhere from being indexed anyway.
|
|
"""
|
|
|
|
from fastapi import FastAPI
|
|
from fastapi.responses import PlainTextResponse
|
|
|
|
from app.core.config import settings
|
|
|
|
NOINDEX_HEADER = (b"x-robots-tag", b"noindex")
|
|
|
|
|
|
def api_robots_txt() -> str:
|
|
"""The API origin's robots.txt body.
|
|
|
|
Disallows everything — except, under path dispatch, the help center. That
|
|
mode (the self-host default) serves public help-center pages from
|
|
{BACKEND_URL}/help/{slug} on this very origin, and those are meant to be
|
|
indexed; a bare "Disallow: /" would quietly de-index every self-hosted
|
|
help center. Cloud runs subdomain mode, where nothing on this origin is a
|
|
page and the carve-out is absent.
|
|
|
|
The Allow goes FIRST. RFC 9309 crawlers take the most specific match and
|
|
would get this right either way, but ones that take the first match read
|
|
a leading "Disallow: /" and stop — skipping the very pages the carve-out
|
|
exists to keep crawlable.
|
|
"""
|
|
lines = ["User-agent: *"]
|
|
if settings.HELP_CENTER_PUBLIC_MODE == "path":
|
|
lines.append("Allow: /help/")
|
|
lines.append("Disallow: /")
|
|
return "\n".join(lines) + "\n"
|
|
|
|
|
|
class NoIndexHeaderMiddleware:
|
|
"""Stamp ``X-Robots-Tag: noindex`` on every response under ``prefix``.
|
|
|
|
Pure ASGI rather than BaseHTTPMiddleware: this fronts every API request,
|
|
and BaseHTTPMiddleware would pull streaming responses through a queue for
|
|
the sake of one header.
|
|
"""
|
|
|
|
def __init__(self, app, prefix: str):
|
|
self.app = app
|
|
self.prefix = prefix
|
|
|
|
async def __call__(self, scope, receive, send):
|
|
if scope["type"] != "http" or not scope.get("path", "").startswith(self.prefix):
|
|
await self.app(scope, receive, send)
|
|
return
|
|
|
|
async def send_with_noindex(message):
|
|
if message["type"] == "http.response.start":
|
|
message["headers"] = [*(message.get("headers") or []), NOINDEX_HEADER]
|
|
await send(message)
|
|
|
|
await self.app(scope, receive, send_with_noindex)
|
|
|
|
|
|
def install_crawler_policy(app: FastAPI) -> None:
|
|
"""Register the robots.txt route and the noindex header middleware."""
|
|
|
|
@app.get("/robots.txt", response_class=PlainTextResponse, include_in_schema=False)
|
|
async def robots_txt():
|
|
return PlainTextResponse(api_robots_txt())
|
|
|
|
app.add_middleware(NoIndexHeaderMiddleware, prefix=settings.API_V1_STR)
|