Create images.py
Browse files- gemini_client/images.py +202 -0
gemini_client/images.py
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# -*- coding: utf-8 -*-
|
| 2 |
+
import os
|
| 3 |
+
import random
|
| 4 |
+
import re
|
| 5 |
+
from datetime import datetime
|
| 6 |
+
from pathlib import Path
|
| 7 |
+
from typing import Dict, Union, Optional
|
| 8 |
+
|
| 9 |
+
from pydantic import BaseModel, field_validator
|
| 10 |
+
from curl_cffi import CurlError
|
| 11 |
+
from curl_cffi.requests import AsyncSession
|
| 12 |
+
from requests.exceptions import HTTPError, RequestException # Ensure RequestException is imported
|
| 13 |
+
|
| 14 |
+
from rich.console import Console
|
| 15 |
+
|
| 16 |
+
console = Console() # Instantiate console for logging
|
| 17 |
+
|
| 18 |
+
class Image(BaseModel):
|
| 19 |
+
"""
|
| 20 |
+
Represents a single image object returned from Gemini.
|
| 21 |
+
|
| 22 |
+
Attributes:
|
| 23 |
+
url (str): URL of the image.
|
| 24 |
+
title (str): Title of the image (default: "[Image]").
|
| 25 |
+
alt (str): Optional description of the image.
|
| 26 |
+
proxy (str | dict | None): Proxy used when saving the image.
|
| 27 |
+
impersonate (str): Browser profile for curl_cffi to impersonate.
|
| 28 |
+
"""
|
| 29 |
+
url: str
|
| 30 |
+
title: str = "[Image]"
|
| 31 |
+
alt: str = ""
|
| 32 |
+
proxy: Optional[Union[str, Dict[str, str]]] = None
|
| 33 |
+
impersonate: str = "chrome110"
|
| 34 |
+
|
| 35 |
+
def __str__(self):
|
| 36 |
+
return f"{self.title}({self.url}) - {self.alt}"
|
| 37 |
+
|
| 38 |
+
def __repr__(self):
|
| 39 |
+
short_url = self.url if len(self.url) <= 50 else self.url[:20] + "..." + self.url[-20:]
|
| 40 |
+
short_alt = self.alt[:30] + "..." if len(self.alt) > 30 else self.alt
|
| 41 |
+
return f"Image(title='{self.title}', url='{short_url}', alt='{short_alt}')"
|
| 42 |
+
|
| 43 |
+
async def save(
|
| 44 |
+
self,
|
| 45 |
+
path: str = "downloaded_images",
|
| 46 |
+
filename: Optional[str] = None,
|
| 47 |
+
cookies: Optional[dict] = None,
|
| 48 |
+
verbose: bool = False,
|
| 49 |
+
skip_invalid_filename: bool = True,
|
| 50 |
+
) -> Optional[str]:
|
| 51 |
+
"""
|
| 52 |
+
Save the image to disk using curl_cffi.
|
| 53 |
+
Parameters:
|
| 54 |
+
path: str, optional
|
| 55 |
+
Directory to save the image (default "downloaded_images").
|
| 56 |
+
filename: str, optional
|
| 57 |
+
Filename to use; if not provided, inferred from URL.
|
| 58 |
+
cookies: dict, optional
|
| 59 |
+
Cookies used for the image request.
|
| 60 |
+
verbose: bool, optional
|
| 61 |
+
If True, outputs status messages (default False).
|
| 62 |
+
skip_invalid_filename: bool, optional
|
| 63 |
+
If True, skips saving if the filename is invalid.
|
| 64 |
+
Returns:
|
| 65 |
+
Absolute path of the saved image if successful; None if skipped.
|
| 66 |
+
Raises:
|
| 67 |
+
HTTPError if the network request fails.
|
| 68 |
+
RequestException/CurlError for other network errors.
|
| 69 |
+
IOError if file writing fails.
|
| 70 |
+
"""
|
| 71 |
+
# Generate filename from URL if not provided
|
| 72 |
+
if not filename:
|
| 73 |
+
try:
|
| 74 |
+
from urllib.parse import urlparse, unquote
|
| 75 |
+
parsed_url = urlparse(self.url)
|
| 76 |
+
base_filename = os.path.basename(unquote(parsed_url.path))
|
| 77 |
+
# Remove invalid characters for filenames
|
| 78 |
+
safe_filename = re.sub(r'[<>:"/\|?*]', '_', base_filename)
|
| 79 |
+
if safe_filename and len(safe_filename) > 0:
|
| 80 |
+
filename = safe_filename
|
| 81 |
+
else:
|
| 82 |
+
filename = f"image_{random.randint(1000, 9999)}.jpg"
|
| 83 |
+
except Exception:
|
| 84 |
+
filename = f"image_{random.randint(1000, 9999)}.jpg"
|
| 85 |
+
|
| 86 |
+
# Validate filename length
|
| 87 |
+
try:
|
| 88 |
+
_ = Path(filename)
|
| 89 |
+
max_len = 255
|
| 90 |
+
if len(filename) > max_len:
|
| 91 |
+
name, ext = os.path.splitext(filename)
|
| 92 |
+
filename = name[:max_len - len(ext) - 1] + ext
|
| 93 |
+
except (OSError, ValueError):
|
| 94 |
+
if verbose:
|
| 95 |
+
console.log(f"[yellow]Invalid filename generated: {filename}[/yellow]")
|
| 96 |
+
if skip_invalid_filename:
|
| 97 |
+
if verbose:
|
| 98 |
+
console.log("[yellow]Skipping save due to invalid filename.[/yellow]")
|
| 99 |
+
return None
|
| 100 |
+
filename = f"image_{random.randint(1000, 9999)}.jpg"
|
| 101 |
+
if verbose:
|
| 102 |
+
console.log(f"[yellow]Using fallback filename: {filename}[/yellow]")
|
| 103 |
+
|
| 104 |
+
# Prepare proxy dictionary for curl_cffi
|
| 105 |
+
proxies_dict = None
|
| 106 |
+
if isinstance(self.proxy, str):
|
| 107 |
+
proxies_dict = {"http": self.proxy, "https": self.proxy}
|
| 108 |
+
elif isinstance(self.proxy, dict):
|
| 109 |
+
proxies_dict = self.proxy
|
| 110 |
+
|
| 111 |
+
try:
|
| 112 |
+
# Use AsyncSession from curl_cffi
|
| 113 |
+
async with AsyncSession(
|
| 114 |
+
cookies=cookies,
|
| 115 |
+
proxies=proxies_dict,
|
| 116 |
+
impersonate=self.impersonate
|
| 117 |
+
# follow_redirects is handled automatically by curl_cffi
|
| 118 |
+
) as client:
|
| 119 |
+
if verbose:
|
| 120 |
+
console.log(f"Attempting to download image from: {self.url}")
|
| 121 |
+
|
| 122 |
+
response = await client.get(self.url)
|
| 123 |
+
response.raise_for_status()
|
| 124 |
+
|
| 125 |
+
# Check content type
|
| 126 |
+
content_type = response.headers.get("content-type", "").lower()
|
| 127 |
+
if "image" not in content_type and verbose:
|
| 128 |
+
console.log(f"[yellow]Warning: Content type is '{content_type}', not an image. Saving anyway.[/yellow]")
|
| 129 |
+
|
| 130 |
+
# Create directory and save file
|
| 131 |
+
dest_path = Path(path)
|
| 132 |
+
dest_path.mkdir(parents=True, exist_ok=True)
|
| 133 |
+
dest = dest_path / filename
|
| 134 |
+
|
| 135 |
+
# Write image data to file
|
| 136 |
+
dest.write_bytes(response.content)
|
| 137 |
+
|
| 138 |
+
if verbose:
|
| 139 |
+
console.log(f"Image saved successfully as {dest.resolve()}")
|
| 140 |
+
|
| 141 |
+
return str(dest.resolve())
|
| 142 |
+
|
| 143 |
+
except HTTPError as e:
|
| 144 |
+
console.log(f"[red]Error downloading image {self.url}: {e.response.status_code} {e}[/red]")
|
| 145 |
+
raise
|
| 146 |
+
except (RequestException, CurlError) as e:
|
| 147 |
+
console.log(f"[red]Network error downloading image {self.url}: {e}[/red]")
|
| 148 |
+
raise
|
| 149 |
+
except IOError as e:
|
| 150 |
+
console.log(f"[red]Error writing image file to {dest}: {e}[/red]")
|
| 151 |
+
raise
|
| 152 |
+
except Exception as e:
|
| 153 |
+
console.log(f"[red]An unexpected error occurred during image save: {e}[/red]")
|
| 154 |
+
raise
|
| 155 |
+
|
| 156 |
+
|
| 157 |
+
class WebImage(Image):
|
| 158 |
+
"""
|
| 159 |
+
Represents an image retrieved from web search results.
|
| 160 |
+
|
| 161 |
+
Returned when asking Gemini to "SEND an image of [something]".
|
| 162 |
+
"""
|
| 163 |
+
pass
|
| 164 |
+
|
| 165 |
+
class GeneratedImage(Image):
|
| 166 |
+
"""
|
| 167 |
+
Represents an image generated by Google's AI image generator (e.g., ImageFX).
|
| 168 |
+
|
| 169 |
+
Attributes:
|
| 170 |
+
cookies (dict[str, str]): Cookies required for accessing the generated image URL,
|
| 171 |
+
typically from the GeminiClient/Chatbot instance.
|
| 172 |
+
"""
|
| 173 |
+
cookies: Dict[str, str]
|
| 174 |
+
|
| 175 |
+
# Updated validator for Pydantic V2
|
| 176 |
+
@field_validator("cookies")
|
| 177 |
+
@classmethod
|
| 178 |
+
def validate_cookies(cls, v: Dict[str, str]) -> Dict[str, str]:
|
| 179 |
+
"""Ensures cookies are provided for generated images."""
|
| 180 |
+
if not v or not isinstance(v, dict):
|
| 181 |
+
raise ValueError("GeneratedImage requires a dictionary of cookies from the client.")
|
| 182 |
+
return v
|
| 183 |
+
|
| 184 |
+
async def save(self, **kwargs) -> Optional[str]:
|
| 185 |
+
"""
|
| 186 |
+
Save the generated image to disk.
|
| 187 |
+
Parameters:
|
| 188 |
+
filename: str, optional
|
| 189 |
+
Filename to use. If not provided, a default name including
|
| 190 |
+
a timestamp and part of the URL is used. Generated images
|
| 191 |
+
are often in .png or .jpg format.
|
| 192 |
+
Additional arguments are passed to Image.save.
|
| 193 |
+
Returns:
|
| 194 |
+
Absolute path of the saved image if successful, None if skipped.
|
| 195 |
+
"""
|
| 196 |
+
if "filename" not in kwargs:
|
| 197 |
+
ext = ".jpg" if ".jpg" in self.url.lower() else ".png"
|
| 198 |
+
url_part = self.url.split('/')[-1][:10]
|
| 199 |
+
kwargs["filename"] = f"{datetime.now().strftime('%Y%m%d%H%M%S')}_{url_part}{ext}"
|
| 200 |
+
|
| 201 |
+
# Pass the required cookies and other args (like impersonate) to the parent save method
|
| 202 |
+
return await super().save(cookies=self.cookies, **kwargs)
|