dan/ocr/document/train.py → dan/ocr/train.py

@@ -10,11 +10,11 @@ import torch
 import torch.multiprocessing as mp
 from torch.optim import Adam
 
-from dan.decoder import GlobalHTADecoder
-from dan.encoder import FCN_Encoder
-from dan.manager.training import Manager
-from dan.mlflow import MLFLOW_AVAILABLE
-from dan.transforms import Preprocessing
+from dan.ocr.decoder import GlobalHTADecoder
+from dan.ocr.encoder import FCN_Encoder
+from dan.ocr.manager.training import Manager
+from dan.ocr.mlflow import MLFLOW_AVAILABLE
+from dan.ocr.transforms import Preprocessing
 from dan.utils import MLflowNotInstalled
 
 if MLFLOW_AVAILABLE:

docs/ref/datasets/analyze/index.md

+# Analysis

docs/ref/datasets/analyze/statistics.md

+# Statistics
+
+::: dan.datasets.analyze.statistics

docs/ref/decoder.md → docs/ref/ocr/decoder.md

 # Decoders
 
-::: dan.decoder
+::: dan.ocr.decoder

docs/ref/ocr/document/index.md

-# Document

docs/ref/encoder.md → docs/ref/ocr/encoder.md

 # Encoder
 
-::: dan.encoder
+::: dan.ocr.encoder

docs/ref/managers/dataset.md → docs/ref/ocr/managers/dataset.md

 # Dataset Managers
 
-::: dan.manager.dataset
+::: dan.ocr.manager.dataset

docs/ref/managers/metrics.md → docs/ref/ocr/managers/metrics.md

 # Metrics managers
 
-::: dan.manager.metrics
+::: dan.ocr.manager.metrics

docs/ref/managers/ocr.md → docs/ref/ocr/managers/ocr.md

 # OCR managers
 
-::: dan.manager.ocr
+::: dan.ocr.manager.ocr

docs/ref/managers/training.md → docs/ref/ocr/managers/training.md

 # Training managers
 
-::: dan.manager.training
+::: dan.ocr.manager.training

docs/ref/mlflow.md → docs/ref/ocr/mlflow.md

 # MLflow
 
-::: dan.mlflow
+::: dan.ocr.mlflow

docs/ref/predict/attention.md → docs/ref/ocr/predict/attention.md

 # Attention
 
-::: dan.predict.attention
+::: dan.ocr.predict.attention

docs/ref/ocr/predict/prediction.md

+# Inference
+
+::: dan.ocr.predict.prediction

docs/ref/schedulers.md → docs/ref/ocr/schedulers.md

 # Schedulers
 
-::: dan.schedulers
+::: dan.ocr.schedulers

docs/ref/ocr/document/train.md → docs/ref/ocr/train.md

 # Training
 
-::: dan.ocr.document.train
+::: dan.ocr.train

docs/ref/transforms.md → docs/ref/ocr/transforms.md

 # Transformations
 
-::: dan.transforms
+::: dan.ocr.transforms

docs/ref/predict/prediction.md

-# Inference
-
-::: dan.predict.prediction

docs/usage/datasets/analyze.md

+# Dataset analysis
+
+## Description
+
+Use the `teklia-dan dataset analyze` command to analyze a dataset. This will display statistics in Markdown format.
+
+The available arguments are
+
+| Parameter       | Description                      | Type  | Default |
+| --------------- | -------------------------------- | ----- | ------- |
+| `--labels`      | Path to the `labels.json` file.  | `str` |         |
+| `--tokens`      | Path to the `tokens.yaml` file.  | `str` | `None`  |
+| `--output-file` | Where the summary will be saved. | `str` |         |
+
+## Examples
+
+### Display statistics for an HTR dataset
+
+```shell
+teklia-dan dataset analyze \
+    --labels path/to/dataset/labels.json \
+    --output-file statistics.md
+```
+
+### Display statistics for an HTR-NER dataset
+
+```shell
+teklia-dan dataset analyze \
+    --labels path/to/dataset/labels.json \
+    --tokens  path/to/tokens.yaml \
+    --output-file statistics.md
+```