Checking in my changes.

.gitignore

@@ -1,3 +1,5 @@
+layoutex_env/*
+src/layoutex/generated-OUTLINE/*
 # Temporary and binary files
 *~
 *.py[cod]
@@ -63,3 +65,4 @@ venv/
 *.idea
 *.pydevproject
 
+layoutex_env/bin/activate

README.md

@@ -0,0 +1,31 @@
+# layoutex
+
+### Requirements
+
+Download and decompress the PubLayNet `labels.tar.gz` here: https://dax-cdn.cdn.appdomain.cloud/dax-publaynet/1.0.0/labels.tar.gz
+
+You can find all of the PubLayNet datasets here: https://developer.ibm.com/exchanges/data/all/publaynet/
+
+Move the publaynet directory into `~/datasets/publaynet`, make a directory called `annotations` and copy the `val.json` into it.
+
+### Running layoutex
+
+To run, do the following:
+
+```bash
+# create the virtual environment
+python3 -m venv layoutex_env
+
+# activate it
+source layoutex_env/bin/activate
+
+# install requirements
+pip install -e .
+pip install -r requirements.txt
+
+# run pytest
+pytest -s tests/test_document_generator.py
+
+# deactivate the venv when finished
+deactivate
+```

requirements.txt

@@ -0,0 +1,21 @@
+python_barcode==0.14.0
+codetiming==1.4.0
+Faker==18.5.1
+imgaug==0.4.0
+importlib_metadata==4.6.4
+matplotlib==3.6.3
+numpy==1.24.1
+#numpy==1.21.5
+opencv_python==4.7.0.68
+#Pillow==9.0.1
+Pillow==9.5.0
+qrcode==7.4.2
+resize_image==0.4.0
+seaborn==0.12.2
+torch==1.13.1
+torchvision==0.14.1
+tqdm==4.64.1
+typing_extensions==4.5.0
+wandb==0.15.0
+pytest==7.3.1
+pytest-asyncio==0.21.0

src/layoutex/component_util.py

@@ -0,0 +1,49 @@
+"""
+Contains document component utility functions.
+"""
+
+import logging
+logger = logging.getLogger(__name__)
+
+def estimate_component_sizing(box, target_size, margin_size):
+    """
+    estimate the component sizing based on the bounding box size and the target size of the document including margins
+    """
+    logger.debug(f"estimate_component_sizing(target_size={target_size}, margin_size={margin_size})")
+    x1, y1, x2, y2 = box
+    target_size = target_size - 2 * margin_size
+    logger.debug(f"computed target_size = {target_size}")
+    w = int(x2 - x1)
+    h = int(y2 - y1)
+    logger.debug(f"computed dimensions  = (w={w}, h={h})")
+    ratio_w = w / target_size
+    ratio_h = h / target_size
+    logger.debug(f"computed ratios      = (ratio_w={ratio_w}, ratio_h={ratio_h})")
+
+    component_w = "FULL_WIDTH"
+    component_h = "FULL_HEIGHT"
+
+    if ratio_w > 0.75:
+        component_w = "FULL_WIDTH"
+    elif ratio_w > 0.5:
+        component_w = "TWO_THIRDS_WIDTH"
+    elif ratio_w > 0.25:
+        component_w = "HALF_WIDTH"
+    elif ratio_w > 0.01:
+        component_w = "QUARTER_WIDTH"
+
+    logger.debug(f"computed component_w = {component_w}")
+
+    if ratio_h > 0.75:
+        component_h = "FULL_HEIGHT"
+    elif ratio_h > 0.25:
+        component_h = "HALF_HEIGHT"
+    elif ratio_h > 0.05:
+        component_h = "QUARTER_HEIGHT"
+    elif ratio_h > 0.01:
+        component_h = "LINE_HEIGHT"
+
+    logger.debug(f"computed component_h = {component_h}")
+
+    # logger.debug(f"estimated component sizing - ratio_w: {ratio_w}, ratio_h: {ratio_h}  -> {component_w}, {component_h}")
+    return component_w, component_h

src/layoutex/config.py

@@ -10,13 +10,13 @@
 noisy_dir = 'image'
 debug_dir = 'debug'
 
-patch_dir = '/home/greg/datasets/dataset/rms/pix2pix-overlay'
-asset_dir = '/home/greg/datasets/dataset/rms/assets'
+patch_dir = '/home/sstauffer/specops/gitlab/layoutex/assets/pix2pix-overlay'
+asset_dir = '/home/sstauffer/specops/gitlab/layoutex/assets'
 
-txt_file_dir = 'text.txt'
+txt_file_dir = '/home/sstauffer/specops/words/words_alpha.txt'
 
 # maximun number of synthetic words to generate
-num_synthetic_imgs = 3000
+num_synthetic_imgs = 100
 train_percentage = 0.8
 
 test_dir = os.path.join(data_dir, val_dir, noisy_dir)