File size: 9,571 Bytes
dc4e6da | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 | """
TODO: latent diffusion model inference
"""
import pathlib
from docgenie.generation.utils.log import log_pipeline_level
from docgenie.generation.utils.stamp import (
create_stamp,
)
import json
from docgenie import ENV
import random
from pathlib import Path
from PIL import Image
import io
from barcode import Code128
from barcode.writer import ImageWriter
from docgenie.generation.models import (
DocLogKey,
PipelineParameters,
SyntheticDatasetFileStructure,
SynDatasetDefinition,
LLMType,
)
from docgenie.generation.utils.status import get_progress_bar
__LOGO_PREFABS__ = ENV.VISUAL_ELEMENT_PREFABS_DIR / "logo"
__FIGURE_PREFABS__ = ENV.VISUAL_ELEMENT_PREFABS_DIR / "figure"
__PHOTO_PREFABS__ = ENV.VISUAL_ELEMENT_PREFABS_DIR / "photo"
_LOGO_CACHE = None
_PHOTO_CACHE = None
_CHART_CACHE = None
def _get_prefabs_paths(image_type: str) -> list[Path]:
"""Cache logo paths to avoid repeated directory scans."""
global _LOGO_CACHE, _PHOTO_CACHE, _CHART_CACHE
image_type_lower = image_type.lower()
if image_type_lower == "logo":
if _LOGO_CACHE is None:
_LOGO_CACHE = _scan_directory(__LOGO_PREFABS__, "logo")
return _LOGO_CACHE
elif image_type_lower == "photo":
if _PHOTO_CACHE is None:
_PHOTO_CACHE = _scan_directory(__PHOTO_PREFABS__, "photo")
return _PHOTO_CACHE
elif image_type_lower == "figure":
if _CHART_CACHE is None:
_CHART_CACHE = _scan_directory(__FIGURE_PREFABS__, "figure")
return _CHART_CACHE
else:
raise ValueError(
f"Invalid image_type: {image_type}. Must be 'logo', 'photo', or 'figure'"
)
def _scan_directory(directory, image_type):
"""Helper to scan directory for images."""
paths = []
for ext in ("*.png", "*.jpg", "*.jpeg"):
paths.extend(directory.glob(ext))
if not paths:
raise FileNotFoundError(f"No {image_type} images found in {directory}")
return paths
"""
{
"id": "ve0",
"type": "stamp",
"type_unmapped": "stamp",
"content": "CONFIDENTIAL",
"rect": {
"x": 766.7671508789062,
"y": 100.63824462890625,
"width": 138.8602294921875,
"height": 138.8602294921875
},
"rotation": -15.0,
"error": null
}
"""
def _prepare_stamp(
result_path: Path, ved: dict, docid: str, dsfiles: SyntheticDatasetFileStructure
):
content = ved["content"]
rotation = ved["rotation"]
width = ved["rect"]["width"]
height = ved["rect"]["height"]
# we dont pass rotation here, each stamp has a slight random rotation, we apply rotation in insertion
stamp = create_stamp(text=content, width=width, height=height, rot_angle=None)
stamp.save(result_path)
def _prepare_logo(
result_path: Path, ved: dict, docid: str, dsfiles: SyntheticDatasetFileStructure
):
logo_paths = _get_prefabs_paths("logo") # getting chached logo paths here
selected_logo_image_path = random.choice(logo_paths)
logo_image = Image.open(selected_logo_image_path).convert(
"RGBA"
) # check this conversion if face any issues
"""If anyone want to do any processing on image do it here->like text insertion"""
logo_image.save(result_path)
# Generate barcode with transparent background
writer = ImageWriter()
writer.set_options(
{ # I think we have to play around with these numbers
"module_width": 0.3,
"module_height": 15.0,
"quiet_zone": 6.5,
"font_size": 7,
"text_distance": 5,
"background": "rgba(255, 255, 255, 0)", # Transparent background
"foreground": "black",
}
)
def _prepare_barcode(
result_path: Path, ved: dict, docid: str, dsfiles: SyntheticDatasetFileStructure
):
content = ved["content"]
if content and content.strip().isdigit():
barcode_content = content.strip()
else:
# Generate random number if content is invalid or empty
barcode_content = str(
random.randint(100000000000, 999999999999)
) # 12-digit number
code128 = Code128(barcode_content, writer=writer)
# Save to buffer first to handle transparency
buffer = io.BytesIO()
code128.write(buffer, options={"format": "PNG"})
buffer.seek(0)
barcode_image = Image.open(buffer).convert("RGBA") # Transparent background
barcode_image.save(result_path)
def _prepare_photo(
result_path: Path, ved: dict, docid: str, dsfiles: SyntheticDatasetFileStructure
):
photo_paths = _get_prefabs_paths("photo") # getting chached photo paths here
selected_photo_image_path = random.choice(photo_paths)
photo_image = Image.open(
selected_photo_image_path
) # check this conversion if face any issues
photo_image.save(result_path)
def _prepare_figure(
result_path: Path, ved: dict, docid: str, dsfiles: SyntheticDatasetFileStructure
):
chart_paths = _get_prefabs_paths("figure") # getting chached charts paths here
selected_chart_image_path = random.choice(chart_paths)
chart_image = Image.open(
selected_chart_image_path
) # check this conversion if face any issues
chart_image.save(result_path)
def process_visual_element_definition(
ved: dict, docid: str, dsfiles: SyntheticDatasetFileStructure
) -> dict:
content = ved["content"]
ved_id = ved["id"]
error = ved["error"]
log = {
"id": ved_id,
"type": ved["type"],
"type_unmapped": ved["type_unmapped"],
"content": content,
"error": error,
}
document_visual_elements_dir = dsfiles.visual_elements_directory / docid
document_visual_elements_dir.mkdir(parents=True, exist_ok=True)
result_path = document_visual_elements_dir / f"{ved_id}.png"
# Skip already generated vis elements
if error is None and not result_path.exists():
match ved["type"]:
case "stamp":
_prepare_stamp(
result_path=result_path,
ved=ved,
docid=docid,
dsfiles=dsfiles,
)
case "logo":
_prepare_logo(
result_path=result_path,
ved=ved,
docid=docid,
dsfiles=dsfiles,
)
case "barcode":
_prepare_barcode(
result_path=result_path,
ved=ved,
docid=docid,
dsfiles=dsfiles,
)
case "photo":
_prepare_photo(
result_path=result_path,
ved=ved,
docid=docid,
dsfiles=dsfiles,
)
case "figure":
_prepare_figure(
result_path=result_path,
ved=ved,
docid=docid,
dsfiles=dsfiles,
)
case _:
log["error"] = "unknown-type"
log["image_path"] = str(result_path) if result_path is not None else None
return log
def prepare_visual_elements(
defs: list[dict], docid: str, dsfiles: SyntheticDatasetFileStructure
) -> list[dict]:
logs = []
random.seed(docid)
for ved in defs:
log = process_visual_element_definition(ved, docid=docid, dsfiles=dsfiles)
logs.append(log)
return logs
def pipeline_create_visual_elements(params: PipelineParameters):
log_pipeline_level()
dsdef = params.dsdef
dsfiles = dsdef.get_file_structure()
# Get valid documents
valid_documents = []
total_pdfs_count = 0
for doclog in dsdef.get_document_logs():
total_pdfs_count += 1
if doclog.pdf_num_pages == 1:
has_visual_elements = doclog.visual_elements_num_elements > 0
if has_visual_elements:
valid_documents.append(doclog.document_id)
print(
f"{len(valid_documents)} of {total_pdfs_count} documents valid for visual element generation."
)
with get_progress_bar() as progress:
insert_task = progress.add_task(
"[red]Creating visual elements...", total=len(valid_documents)
)
for docid in valid_documents:
visual_element_def_file = (
dsfiles.visual_element_definitions_directory / f"{docid}.json"
)
visual_element_definitions = json.loads(
visual_element_def_file.read_text(encoding="utf-8")
)
insertion_logs = prepare_visual_elements(
defs=visual_element_definitions, docid=docid, dsfiles=dsfiles
)
errors = [
f"{d['id']}: {d['error']}"
for d in insertion_logs
if d["error"] is not None
]
dsdef.write_to_document_log(
document_id=docid,
vals={
DocLogKey.visual_elements_generation_logs: insertion_logs,
DocLogKey.visual_elements_generation_errors: errors,
},
)
progress.update(insert_task, advance=1)
|