speedy-utils 1.1.27__py3-none-any.whl → 1.1.29__py3-none-any.whl

speedy_utils-1.1.29.dist-info/RECORD

@@ -0,0 +1,57 @@
+llm_utils/__init__.py,sha256=tSo9RfmOhH1ofZseRtwr4iOpIQ1zKRocAzmuy53RgiI,1634
+llm_utils/group_messages.py,sha256=_XuFkEkO_iQDjVVx80XNFpufGH6iIDwHPDoCRJa60Ak,3691
+llm_utils/chat_format/__init__.py,sha256=a7BKtBVktgLMq2Do4iNu3YfdDdTG1v9M_BkmaEorp5c,775
+llm_utils/chat_format/display.py,sha256=Lffjzna9_vV3QgfiXZM2_tuVb3wqA-WxwrmoAjsJigw,17356
+llm_utils/chat_format/transform.py,sha256=PJ2g9KT1GSbWuAs7giEbTpTAffpU9QsIXyRlbfpTZUQ,5351
+llm_utils/chat_format/utils.py,sha256=M2EctZ6NeHXqFYufh26Y3CpSphN0bdZm5xoNaEJj5vg,1251
+llm_utils/lm/__init__.py,sha256=lFE2DZRpj6eRMo11kx7oRLyYOP2FuDmz08mAcq-cYew,730
+llm_utils/lm/base_prompt_builder.py,sha256=_TzYMsWr-SsbA_JNXptUVN56lV5RfgWWTrFi-E8LMy4,12337
+llm_utils/lm/llm.py,sha256=Ryne4why4VgRCvTJW1SLJciv2pcYYROGP0AwkLVdGPg,16299
+llm_utils/lm/llm_signature.py,sha256=vV8uZgLLd6ZKqWbq0OPywWvXAfl7hrJQnbtBF-VnZRU,1244
+llm_utils/lm/lm_base.py,sha256=Bk3q34KrcCK_bC4Ryxbc3KqkiPL39zuVZaBQ1i6wJqs,9437
+llm_utils/lm/mixins.py,sha256=on83g-JO2SpZ0digOpU8mooqFBX6w7Bc-DeGzVoVCX8,14536
+llm_utils/lm/openai_memoize.py,sha256=rYrSFPpgO7adsjK1lVdkJlhqqIw_13TCW7zU8eNwm3o,5185
+llm_utils/lm/signature.py,sha256=K1hvCAqoC5CmsQ0Y_ywnYy2fRb5JzmIK8OS-hjH-5To,9971
+llm_utils/lm/utils.py,sha256=t-RSR-ffOs2HT67JfgqRhbZX9wbxlJNtQMro4ZyuVVM,12461
+llm_utils/lm/async_lm/__init__.py,sha256=j0xK49ooZ0Dm5GstGGHbmPMrPjd3mOXoJ1H7eAL_Z4g,122
+llm_utils/lm/async_lm/_utils.py,sha256=mB-AueWJJatTx0PXqd_oWc6Kz36cfgDmDTKgiXafCJI,6106
+llm_utils/lm/async_lm/async_llm_task.py,sha256=2PWW4vPW2jYUiGmYFo4-DHrmX5Jm8Iw_1qo6EPL-ytE,18611
+llm_utils/lm/async_lm/async_lm.py,sha256=W8n_S5PKJln9bzO9T525-tqo5DFwBZNXDucz_v-35iU,13761
+llm_utils/lm/async_lm/async_lm_base.py,sha256=ga5nCzows5Ye3yop41zsUxNYxcj_Vpf02DsfJ1eoE9U,8358
+llm_utils/lm/async_lm/lm_specific.py,sha256=PxP54ltrh9NrLJx7BPib52oYo_aCvDOjf7KzMjp1MYg,3929
+llm_utils/scripts/README.md,sha256=yuOLnLa2od2jp4wVy3rV0rESeiV3o8zol5MNMsZx0DY,999
+llm_utils/scripts/vllm_load_balancer.py,sha256=eQlH07573EDWIBkwc9ef1WvI59anLr4hQqLfZvQk7xk,37133
+llm_utils/scripts/vllm_serve.py,sha256=tPcRB_MbJ01LcqC83RHQ7W9XDS7b1Ntc0fCRdegsNXU,14761
+llm_utils/vector_cache/__init__.py,sha256=oZXpjgBuutI-Pd_pBNYAQMY7-K2C6xv8Qt6a3p88GBQ,879
+llm_utils/vector_cache/cli.py,sha256=MAvnmlZ7j7_0CvIcSyK4TvJlSRFWYkm4wE7zSq3KR8k,6246
+llm_utils/vector_cache/core.py,sha256=VXuYJy1AX22NHKvIXRriETip5RrmQcNp73-g-ZT774o,30950
+llm_utils/vector_cache/types.py,sha256=CpMZanJSTeBVxQSqjBq6pBVWp7u2-JRcgY9t5jhykdQ,438
+llm_utils/vector_cache/utils.py,sha256=OsiRFydv8i8HiJtPL9hh40aUv8I5pYfg2zvmtDi4DME,1446
+speedy_utils/__imports.py,sha256=KQogps2TTzigJK8YkW875qaIEwf8PypmTtzyc7svvfw,5691
+speedy_utils/__init__.py,sha256=dxEwqFTQIXIsWlKEXm6b5pcEx2c8W8R5iblbHN2chdc,2656
+speedy_utils/all.py,sha256=5gN_mIvx9mtEWMXl64S0NvY3Wrj1bb5QyrMh-U03uiE,2508
+speedy_utils/common/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+speedy_utils/common/clock.py,sha256=raLtMGIgzrRej5kUt7hOUm2ZZw2THVPo-q8dMvdZOxw,7354
+speedy_utils/common/function_decorator.py,sha256=GKXqRs_hHFFmhyhql0Br0o52WzekUnpNlm99NfaVwgY,2025
+speedy_utils/common/logger.py,sha256=pBZwLDduoVzXd1o-hR43Pkdt7farnjS-yZOFefAw5eA,6275
+speedy_utils/common/notebook_utils.py,sha256=6mxXZcTHwYob3nobAzbSZnDyXRahFaaSko1_SYlvZy4,1905
+speedy_utils/common/patcher.py,sha256=Rku-N4DJNue8BCLUx7y3ad_3t_WU2HleHKlbR0vhaRc,2319
+speedy_utils/common/report_manager.py,sha256=78KQ0gUzvbzT6EjHYZ5zgtV41cPRvdX8hnF2oSWA4qA,3849
+speedy_utils/common/utils_cache.py,sha256=1UAqOSb4nFVlhuQRfTEXCN-8Wf6yntXyMA6yp61-98I,26277
+speedy_utils/common/utils_io.py,sha256=w9AxMD_8V3Wyo_0o9OtXjVQS8Z3KhxQiOkrl2p8Np70,15762
+speedy_utils/common/utils_misc.py,sha256=ZRJCS7OJxybpVm1sasoeCYRW2TaaGCXj4DySYlQeVR8,2227
+speedy_utils/common/utils_print.py,sha256=AGDB7mgJnO00QkJBH6kJb46738q3GzMUZPwtQ248vQw,4763
+speedy_utils/multi_worker/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+speedy_utils/multi_worker/process.py,sha256=O3BpGH7iL_2vh_ezwyHb28lvkVADABpTUnhKHbiEe8I,10542
+speedy_utils/multi_worker/thread.py,sha256=k4Ff4R2W0Ehet1zJ5nHQOfcsvOjnJzU6A2I18qw7_6M,21320
+speedy_utils/scripts/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+speedy_utils/scripts/mpython.py,sha256=aZvusJLKa3APVhabbFUAEo873VBm8Fym7HKGmVW4LyE,3843
+speedy_utils/scripts/openapi_client_codegen.py,sha256=GModmmhkvGnxljK4KczyixKDrk-VEcLaW5I0XT6tzWo,9657
+vision_utils/README.md,sha256=AIDZZj8jo_QNrEjFyHwd00iOO431s-js-M2dLtVTn3I,5740
+vision_utils/__init__.py,sha256=XsLxy1Fn33Zxu6hTFl3NEWfxGjuQQ-0Wmoh6lU9NZ_o,257
+vision_utils/io_utils.py,sha256=1FkG6k7uwZALh3-JkWXEHoGQJhjTqG1jC20SxObPRS0,25921
+vision_utils/plot.py,sha256=3NW2atYPdcs6-Qq8N179Pwji-p5ZUN6ddC0zE2WZb_8,11984
+speedy_utils-1.1.29.dist-info/METADATA,sha256=FcgtzlK1CJlS3sWFYi3Apkx4qqTcTRoLOaHz6YW8Wuc,8028
+speedy_utils-1.1.29.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+speedy_utils-1.1.29.dist-info/entry_points.txt,sha256=1rrFMfqvaMUE9hvwGiD6vnVh98kmgy0TARBj-v0Lfhs,244
+speedy_utils-1.1.29.dist-info/RECORD,,

vision_utils/README.md

@@ -0,0 +1,202 @@
+# Vision Utils
+
+Utility functions for computer vision tasks, particularly for visualizing images in Jupyter notebooks.
+
+## Features
+
+### `plot_images_notebook(images, ...)`
+
+Plot a batch of images in a notebook with automatic format detection, smart grid layout, and size optimization to prevent notebook breaking.
+
+**Supported Input Formats:**
+- **Numpy arrays:**
+  - Batch: `(B, H, W, C)` or `(B, C, H, W)`
+  - Single: `(H, W, C)`, `(C, H, W)`, or `(H, W)` for grayscale
+- **PyTorch tensors:**
+  - Batch: `(B, H, W, C)` or `(B, C, H, W)`
+  - Single: `(H, W, C)`, `(C, H, W)`, or `(H, W)` for grayscale
+- **List of images:**
+  - Mixed formats (numpy arrays and/or torch tensors)
+
+**Parameters:**
+- `images`: Images to plot (see supported formats above)
+- `nrows`: Number of rows in grid (default: None, auto-calculated using sqrt)
+- `ncols`: Number of columns in grid (default: None, auto-calculated using sqrt)
+- `figsize`: Figure size `(width, height)` in inches (default: auto-calculated based on grid size and image count)
+- `titles`: Optional list of titles for each image
+- `cmap`: Colormap for grayscale images (default: 'gray')
+- `dpi`: Dots per inch for the figure (default: 72)
+- `max_figure_width`: Maximum figure width in inches (default: 15.0)
+- `max_figure_height`: Maximum figure height in inches (default: 20.0)
+
+**Smart Features:**
+- **Auto Grid Layout**: When both `nrows` and `ncols` are None, uses sqrt to create roughly square grid
+- **Adaptive Cell Sizing**: Automatically reduces cell size for larger image counts to prevent notebook breaking
+- **Size Constraints**: Respects max width/height limits to ensure plots render properly in notebooks
+- **Aspect Ratio Aware**: Calculates figure size based on actual image aspect ratios
+
+## Installation
+
+The vision_utils module is part of speedy_utils. Install with:
+
+```bash
+pip install -e .
+```
+
+For PyTorch support:
+```bash
+pip install torch
+```
+
+For plotting (required):
+```bash
+pip install matplotlib
+```
+
+## Usage Examples
+
+### Auto Grid Layout (Recommended)
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# Auto grid - creates 3x3 grid for 9 images
+images = np.random.rand(9, 64, 64, 3)
+plot_images_notebook(images)
+
+# Auto grid - creates 3x3 grid for 8 images (1 empty cell)
+images = np.random.rand(8, 64, 64, 3)
+plot_images_notebook(images)
+```
+
+### Manual Grid Layout
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# Specify exact grid dimensions
+images = np.random.rand(8, 64, 64, 3)
+plot_images_notebook(images, nrows=2, ncols=4)
+```
+
+### Basic Usage
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# Batch of images - auto grid layout
+images = np.random.rand(8, 64, 64, 3)
+plot_images_notebook(images)
+```
+
+### Many Images (Adaptive Sizing)
+```
+
+### Many Images (Adaptive Sizing)
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# Automatically uses smaller cells for many images
+images = np.random.rand(25, 64, 64, 3)
+plot_images_notebook(images)  # Creates 5x5 grid with optimized cell size
+```
+
+### PyTorch Tensors
+
+```python
+import torch
+from vision_utils import plot_images_notebook
+
+# PyTorch tensor in (B, C, H, W) format
+images = torch.rand(8, 3, 64, 64)
+plot_images_notebook(images)
+```
+
+### Mixed Formats
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# List of images with different formats
+images = [
+    np.random.rand(64, 64, 3),  # (H, W, C)
+    np.random.rand(3, 64, 64),  # (C, H, W)
+    np.random.rand(64, 64),     # Grayscale
+]
+
+plot_images_notebook(images, titles=titles)
+```
+
+### Single Image
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# Single image
+image = np.random.rand(128, 128, 3)
+plot_images_notebook(image)
+```
+
+### Custom Figure Size and Colormap
+
+```python
+import numpy as np
+from vision_utils import plot_images_notebook
+
+# Grayscale images with custom settings
+images = np.random.rand(4, 64, 64)
+plot_images_notebook(
+    images,
+    nrows=2,
+    ncols=2,
+    figsize=(8, 8),
+    cmap='viridis',
+    dpi=100
+)
+```
+
+## Smart Grid Layout
+
+When `nrows` and `ncols` are both `None` (default), the function automatically calculates the optimal grid layout:
+
+1. Uses `sqrt(n_images)` to determine a roughly square grid
+2. For 9 images → 3x3 grid
+3. For 8 images → 3x3 grid (with 1 empty cell)
+4. For 16 images → 4x4 grid
+5. For 25 images → 5x5 grid
+
+This ensures balanced layouts without manual calculation.
+
+## Adaptive Cell Sizing
+
+The function automatically adjusts cell sizes based on the number of images to prevent notebook breaking:
+
+- **1-4 images**: 4.0 inches per cell
+- **5-9 images**: 3.0 inches per cell
+- **10-16 images**: 2.5 inches per cell
+- **17+ images**: 2.0 inches per cell
+
+Additionally, it respects `max_figure_width` (default: 15") and `max_figure_height` (default: 20") constraints to ensure plots render properly in notebooks.
+
+## Format Detection
+
+The function automatically detects and converts between different image formats:
+
+1. **Channel Detection:** Identifies if channels are in the first (`C, H, W`) or last (`H, W, C`) dimension
+2. **Batch Detection:** Handles both batched and single images
+3. **Type Conversion:** Automatically converts PyTorch tensors to numpy arrays
+4. **Value Normalization:** Handles both `[0, 1]` and `[0, 255]` value ranges
+
+## Notes
+
+- The function assumes that if the first dimension is 1 or 3 and is smaller than the other dimensions, it represents channels (`C, H, W` format)
+- Images with values in `[0, 1]` are displayed directly
+- Images with values > 1 are assumed to be in `[0, 255]` range and normalized to `[0, 1]`
+- Grayscale images can be 2D `(H, W)` or 3D `(H, W, 1)`

vision_utils/__init__.py

@@ -0,0 +1,4 @@
+from .io_utils import read_images, read_images_cpu, read_images_gpu, ImageMmap, ImageMmapDynamic
+from .plot import plot_images_notebook
+
+__all__ = ['plot_images_notebook', 'read_images_cpu', 'read_images_gpu', 'read_images', 'ImageMmap', 'ImageMmapDynamic']