Spaces:
Runtime error
Runtime error
File size: 7,848 Bytes
9e93b10 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 | import asyncio
import base64
import io
import mimetypes
import re
from pathlib import Path
from typing import Optional
from fastapi import (
APIRouter,
Depends,
HTTPException,
Request,
UploadFile,
)
from rexpro_ai.env import (
AIOHTTP_CLIENT_ALLOW_REDIRECTS,
AIOHTTP_CLIENT_SESSION_SSL,
ENABLE_IMAGE_CONTENT_TYPE_EXTENSION_FALLBACK,
)
from rexpro_ai.models.chats import Chats
from rexpro_ai.models.files import Files
from rexpro_ai.retrieval.web.utils import validate_url
from rexpro_ai.routers.files import upload_file_handler
from rexpro_ai.utils.access_control.files import has_access_to_file
from rexpro_ai.routers.images import (
get_image_data,
upload_image,
)
from rexpro_ai.storage.provider import Storage
from rexpro_ai.utils.session_pool import get_session
BASE64_IMAGE_URL_PREFIX = re.compile(r'data:image/\w+;base64,', re.IGNORECASE)
MARKDOWN_IMAGE_URL_PATTERN = re.compile(r'!\[(.*?)\]\((.+?)\)', re.IGNORECASE)
# Extension-based MIME fallback, only used when ENABLE_IMAGE_CONTENT_TYPE_EXTENSION_FALLBACK is True.
_IMAGE_MIME_FALLBACK = {
'.webp': 'image/webp',
'.png': 'image/png',
'.jpg': 'image/jpeg',
'.jpeg': 'image/jpeg',
'.gif': 'image/gif',
'.svg': 'image/svg+xml',
'.bmp': 'image/bmp',
'.tiff': 'image/tiff',
'.tif': 'image/tiff',
'.ico': 'image/x-icon',
'.heic': 'image/heic',
'.heif': 'image/heif',
'.avif': 'image/avif',
}
async def get_image_base64_from_url(url: str, user=None) -> Optional[str]:
try:
if url.startswith('http'):
# Validate URL to prevent SSRF attacks against local/private networks.
# allow_redirects=False prevents redirect-based SSRF: validate_url() is
# called only on the originally-submitted URL; following 3xx redirects
# without re-validation would let an attacker reach private IPs via a
# public host that redirects internally (e.g. cloud-metadata exfil).
validate_url(url)
# Download the image from the URL
session = await get_session()
async with session.get(
url, ssl=AIOHTTP_CLIENT_SESSION_SSL, allow_redirects=AIOHTTP_CLIENT_ALLOW_REDIRECTS
) as response:
response.raise_for_status()
image_data = await response.read()
encoded_string = base64.b64encode(image_data).decode('utf-8')
content_type = response.headers.get('Content-Type', 'image/png')
return f'data:{content_type};base64,{encoded_string}'
else:
# Non-URL string — treat as file_id. Delegate to the canonical
# file-ID resolver which enforces ownership/access checks.
return await get_image_base64_from_file_id(url, user=user)
except Exception as e:
return None
async def get_image_url_from_base64(request, base64_image_string, metadata, user):
if BASE64_IMAGE_URL_PREFIX.match(base64_image_string):
image_url = ''
# Extract base64 image data from the line
image_data, content_type = await get_image_data(base64_image_string)
if image_data is not None:
_, image_url = await upload_image(
request,
image_data,
content_type,
metadata,
user,
)
return image_url
return None
async def convert_markdown_base64_images(request, content: str, metadata, user):
MIN_REPLACEMENT_URL_LENGTH = 1024
result_parts = []
last_end = 0
for match in MARKDOWN_IMAGE_URL_PATTERN.finditer(content):
result_parts.append(content[last_end : match.start()])
base64_string = match.group(2)
if len(base64_string) > MIN_REPLACEMENT_URL_LENGTH:
url = await get_image_url_from_base64(request, base64_string, metadata, user)
if url:
result_parts.append(f'')
else:
result_parts.append(match.group(0))
else:
result_parts.append(match.group(0))
last_end = match.end()
result_parts.append(content[last_end:])
return ''.join(result_parts)
def load_b64_audio_data(b64_str):
try:
if ',' in b64_str:
header, b64_data = b64_str.split(',', 1)
else:
b64_data = b64_str
header = 'data:audio/wav;base64'
audio_data = base64.b64decode(b64_data)
content_type = header.split(';')[0].split(':')[1] if ';' in header else 'audio/wav'
return audio_data, content_type
except Exception as e:
print(f'Error decoding base64 audio data: {e}')
return None, None
async def upload_audio(request, audio_data, content_type, metadata, user):
audio_format = mimetypes.guess_extension(content_type)
file = UploadFile(
file=io.BytesIO(audio_data),
filename=f'generated-{audio_format}', # will be converted to a unique ID on upload_file
headers={
'content-type': content_type,
},
)
file_item = await upload_file_handler(
request,
file=file,
metadata=metadata,
process=False,
user=user,
)
url = request.app.url_path_for('get_file_content_by_id', id=file_item.id)
return url
async def get_audio_url_from_base64(request, base64_audio_string, metadata, user):
if 'data:audio/wav;base64' in base64_audio_string:
audio_url = ''
# Extract base64 audio data from the line
audio_data, content_type = load_b64_audio_data(base64_audio_string)
if audio_data is not None:
audio_url = await upload_audio(
request,
audio_data,
content_type,
metadata,
user,
)
return audio_url
return None
async def get_file_url_from_base64(request, base64_file_string, metadata, user):
if BASE64_IMAGE_URL_PREFIX.match(base64_file_string):
return await get_image_url_from_base64(request, base64_file_string, metadata, user)
elif 'data:audio/wav;base64' in base64_file_string:
return await get_audio_url_from_base64(request, base64_file_string, metadata, user)
return None
async def get_image_base64_from_file_id(id: str, user=None) -> Optional[str]:
file = await Files.get_file_by_id(id)
if not file:
return None
# Gate file-by-id resolution by ownership to prevent exfiltration.
# A caller could place another user's file_id in an image_url field;
# without this check the server reads the file from disk, inlines it
# base64 into the LLM request, and the content leaks via OCR/describe.
# Owner, admin, and explicit read-grant holders are allowed.
if user is None:
return None
if file.user_id != user.id and user.role != 'admin' and not await has_access_to_file(file.id, 'read', user):
return None
try:
file_path = await asyncio.to_thread(Storage.get_file, file.path)
file_path = Path(file_path)
# Check if the file already exists in the cache
if file_path.is_file():
with open(file_path, 'rb') as image_file:
encoded_string = base64.b64encode(image_file.read()).decode('utf-8')
content_type = mimetypes.guess_type(file_path.name)[0] or (file.meta or {}).get('content_type')
if not content_type and ENABLE_IMAGE_CONTENT_TYPE_EXTENSION_FALLBACK:
content_type = _IMAGE_MIME_FALLBACK.get(file_path.suffix.lower())
if not content_type:
return None
return f'data:{content_type};base64,{encoded_string}'
else:
return None
except Exception as e:
return None
|