Trim line lenght in the docuemntation/examples for the new theme

- the new nvidia-theme reccomends the line lenght of the code snippets
  80 characters, not 100 as used in DALI

Signed-off-by: Janusz Lisiecki <jlisiecki@nvidia.com>

.github/workflows/lint.yml

@@ -7,7 +7,8 @@ on:
 # TODO(klecki): Deduplicate this list of directories with `lint.cmake` file
 env:
   PYTHON_SECURITY_LINT_PATHS: "./tools ./dali/python ./dali_tf_plugin"
-  PYTHON_LINT_PATHS: "./dali ./docs ./internal_tools ./qa"
+  PYTHON_LINT_DOCS_PATHS: "./docs"
+  PYTHON_LINT_PATHS: "./dali ./internal_tools ./qa"
   AUTOGRAPH_LINT_PATHS: "./dali/python/nvidia/dali/_autograph ./dali/test/python/autograph/"
 
 jobs:
@@ -20,7 +21,8 @@ jobs:
           python-version: '3.10'
       - run: pip install  flake8 bandit "black[jupyter]"==24.4.2
       - run: black --check --verbose ${{ env.PYTHON_SECURITY_LINT_PATHS }} ${{ env.PYTHON_LINT_PATHS }} ${{ env.AUTOGRAPH_LINT_PATHS }}
-      - run: flake8 --config=.flake8 ${{ env.PYTHON_SECURITY_LINT_PATHS }} ${{ env.PYTHON_LINT_PATHS }}
+      - run: black --check --config pyproject_docs.toml --verbose ${{ env.PYTHON_LINT_DOCS_PATHS }}
+      - run: flake8 --config=.flake8 ${{ env.PYTHON_SECURITY_LINT_PATHS }} ${{ env.PYTHON_LINT_PATHS }} ${{ env.PYTHON_LINT_DOCS_PATHS }}
       - run: flake8 --config=.flake8.ag ${{ env.AUTOGRAPH_LINT_PATHS }}
       - run: bandit --config bandit.yml -r ${{ env.PYTHON_SECURITY_LINT_PATHS }}
   cpp:

cmake/lint.cmake

@@ -25,10 +25,12 @@ set(PYTHON_SECURITY_LINT_PATHS
         ${PROJECT_SOURCE_DIR}/dali_tf_plugin
 )
 
+set (PYTHON_LINT_DOCS_PATHS
+        ${PROJECT_SOURCE_DIR}/docs
+)
 set(PYTHON_LINT_PATHS
         ${PYTHON_SECURITY_LINT_PATHS}
         ${PROJECT_SOURCE_DIR}/dali
-        ${PROJECT_SOURCE_DIR}/docs
         ${PROJECT_SOURCE_DIR}/qa
         ${PROJECT_SOURCE_DIR}/internal_tools
 )
@@ -41,6 +43,8 @@ set(AUTOGRAPH_LINT_PATHS
 add_custom_target(lint-python-black
         COMMAND
           black --check --config ${PROJECT_SOURCE_DIR}/pyproject.toml ${PYTHON_LINT_PATHS} ${AUTOGRAPH_LINT_PATHS}
+        COMMAND
+          black --check --config ${PROJECT_SOURCE_DIR}/pyproject_docs.toml ${PYTHON_LINT_DOCS_PATHS}
         COMMENT
           "Performing black Python formatting check"
 )
@@ -55,7 +59,7 @@ add_custom_target(lint-python-bandit
 
 add_custom_target(lint-python-flake
         COMMAND
-          flake8 --config=${PROJECT_SOURCE_DIR}/.flake8 ${PYTHON_LINT_PATHS}
+          flake8 --config=${PROJECT_SOURCE_DIR}/.flake8 ${PYTHON_LINT_PATHS} ${PYTHON_LINT_DOCS_PATHS} ${PYTHON_LINT_DOCS_PATHS}
         COMMAND
           flake8 --config=${PROJECT_SOURCE_DIR}/.flake8.ag ${AUTOGRAPH_LINT_PATHS}
         COMMENT

dali/operators/decoder/inflate/inflate.cc

@@ -40,11 +40,11 @@ concatenating compressed frames from the corresponding sequences.::
 
   @pipeline_def
   def inflate_sequence_pipeline():
-    compressed_seq, uncompressed_hwc_shape, compressed_chunk_sizes = fn.external_source(...)
+    compres_seq, uncompres_hwc_shape, compres_chunk_sizes = fn.external_source(...)
     sequences = fn.experimental.inflate(
-        compressed_seq.gpu(),
-        chunk_sizes=compressed_chunk_sizes,  # refers to sizes in ``compressed_seq``
-        shape=uncompressed_hwc_shape,
+        compres_seq.gpu(),
+        chunk_sizes=compres_chunk_sizes,  # refers to sizes in ``compres_seq``
+        shape=uncompres_hwc_shape,
         layout="HWC",
         sequence_axis_name="F")
     return sequences

dali/operators/generic/erase/erase.cc

@@ -77,8 +77,11 @@ A fill value is provided for all the channels. The coordinates can be transforme
 multiplying by the input shape.
 What gives::
 
-    output[y, x, c] = 100             if 0.15 * 300 <= x < (0.3 + 0.15) * 300 and 0.15 * 300 <= y < (0.3 + 0.15) * 300
-    output[y, x, c] = input[y, x, c]  otherwise
+    if 0.15 * 300 <= x < (0.3 + 0.15) * 300 and
+       0.15 * 300 <= y < (0.3 + 0.15) * 300:
+      output[y, x, c] = 100
+    else:
+      output[y, x, c] = input[y, x, c]  otherwise
 
 **Example 4:**
 ``anchor`` = (0.15, 0.15), ``shape`` = (20, 30), ``normalized_anchor`` = True, ``normalized_shape`` = False
@@ -90,8 +93,11 @@ coordinates. Since no axis_names is provided, the anchor and shape must contain
 except "C" (channels).
 What gives::
 
-    output[y, x, c] = 0               if 0.15 * 300 <= x < (0.15 * 300) + 20 and (0.15 * 300) <= y < (0.15 * 300) + 30
-    output[y, x, c] = input[y, x, c]  otherwise
+    if 0.15 * 300 <= x < (0.15 * 300) + 20 and
+       (0.15 * 300) <= y < (0.15 * 300) + 30:
+      output[y, x, c] = 0
+    else:
+      output[y, x, c] = input[y, x, c]  otherwise
 )code")
   .NumInput(1)
   .NumOutput(1)

dali/operators/image/color/brightness_contrast.cc

@@ -79,7 +79,8 @@ DALI_SCHEMA(BrightnessContrast)
 
 The brightness and contrast are adjusted based on the following formula::
 
-  out = brightness_shift * output_range + brightness * (contrast_center + contrast * (in - contrast_center))
+  out = brightness_shift * output_range + brightness *
+        (contrast_center + contrast * (in - contrast_center))
 
 Where the output_range is 1 for float outputs or the maximum positive value for integral types.
 

dali/operators/input/video_input_cpu.cc

@@ -60,21 +60,25 @@ will be partial and the last sequence in this batch will be determined using
     -------------------------------------------------------------------
 
 
-    User decided that there shall be 5 frames per sequence and the last_sequence_policy='partial':
+    User decided that there shall be 5 frames per sequence and
+    the last_sequence_policy='partial':
     -------------------------------------------------------------------
     [   ][   ][   ][   ][   ][   ][   ][   ][   ][   ][   ][   ][   ][]
     -------------------------------------------------------------------
-                      Since there are not enough frames, the last sequence comprises 2 frames.
+    Since there are not enough frames, the last sequence comprises 2 frames.
 
 
-    The Pipeline has max_batch_size=3, therefore the operator will return 5 batches of sequences.
-    First 4 batches comprise 3 sequences and the last batch is partial and comprises 2 sequences.
+    The Pipeline has max_batch_size=3, therefore the operator will return
+    5 batches of sequences.
+    First 4 batches comprise 3 sequences and the last batch is partial and
+    comprises 2 sequences.
     ---------------   ---------------   ---------------   ---------------   -------
     [   ][   ][   ]   [   ][   ][   ]   [   ][   ][   ]   [   ][   ][   ]   [   ][]
     ---------------   ---------------   ---------------   ---------------   -------
 
 
-    With the last_sequence_policy='pad', the last sequence of the last batch will be padded with 0:
+    With the last_sequence_policy='pad', the last sequence of the last batch
+    will be padded with 0:
     ---------------   ---------------   ---------------   ---------------   -------000
     [   ][   ][   ]   [   ][   ][   ]   [   ][   ][   ]   [   ][   ][   ]   [   ][   ]
     ---------------   ---------------   ---------------   ---------------   -------000

dali/operators/reader/coco_reader_op.cc

@@ -41,7 +41,8 @@ images and annotation JSON files.
 
 This readers produces the following outputs::
 
-    images, bounding_boxes, labels, ((polygons, vertices) | (pixelwise_masks)), (image_ids)
+    images, bounding_boxes, labels, ((polygons, vertices) | (pixelwise_masks)),
+    (image_ids)
 
 * **images**
   Each sample contains image data with layout ``HWC`` (height, width, channels).

dali/operators/reader/nemo_asr_reader_op.cc

@@ -38,9 +38,22 @@ NVIDIA NeMo compatible manifest.
 
 Example manifest file::
 
-    {"audio_filepath": "path/to/audio1.wav", "duration": 3.45, "text": "this is a nemo tutorial"}
-    {"audio_filepath": "path/to/audio1.wav", "offset": 3.45, "duration": 1.45, "text": "same audio file but using offset"}
-    {"audio_filepath": "path/to/audio2.wav", "duration": 3.45, "text": "third transcript in this example"}
+    {
+      "audio_filepath": "path/to/audio1.wav",
+      "duration": 3.45,
+      "text": "this is a nemo tutorial"
+    }
+    {
+      "audio_filepath": "path/to/audio1.wav",
+      "offset": 3.45,
+      "duration": 1.45,
+      "text": "same audio file but using offset"
+    }
+    {
+      "audio_filepath": "path/to/audio2.wav",
+      "duration": 3.45,
+      "text": "third transcript in this example"
+    }
 
 .. note::
     Only ``audio_filepath`` is field mandatory. If ``duration`` is not specified, the whole audio file will be used. A missing ``text`` field

dali/operators/segmentation/select_masks.cc

@@ -40,7 +40,8 @@ masks are present.
 Let us assume the following input mask, where symbolic coordinates are used for a clearer example::
 
     polygons = [[0, 0, 3], [1, 3, 7], [2, 7, 10]]
-    vertices = [[x0, y0], [x1, y1], [x2, y2], [x3, y3], [x4, y4], [x5, y5], [x6, y6], [x7, y7], [x8, y8], [x9, y9]]
+    vertices = [[x0, y0], [x1, y1], [x2, y2], [x3, y3], [x4, y4], [x5, y5],
+                [x6, y6], [x7, y7], [x8, y8], [x9, y9]]
 
 Example 1: Selecting a single mask with id ``1``, maintaining the original id::
 

dali/python/nvidia/dali/plugin/mxnet.py

@@ -481,7 +481,8 @@ class DALIClassificationIterator(DALIGenericIterator):
 
     .. code-block:: python
 
-       DALIClassificationIterator(pipelines, reader_name, data_name, label_name, data_layout)
+       DALIClassificationIterator(pipelines, reader_name, data_name, label_name,
+                                  data_layout)
 
     is equivalent to calling
 

docs/advanced_topics_sharding.rst

@@ -62,12 +62,14 @@ Shard calculation
 
 Here is the formula to calculate the shard size for a shard ID::
 
-    floor((id + 1) * dataset_size / num_shards) - floor(id * dataset_size / num_shards)
+    floor((id + 1) * dataset_size / num_shards) -
+        floor(id * dataset_size / num_shards)
 
 When the pipeline advances through the epochs and the reader moves to the next shard, the formula
 needs to be extended to reflect this change::
 
-  floor(((id + epoch_num) % num_shards + 1) * dataset_size / num_shards) - floor(((id + epoch_num) % num_shards) * dataset_size / num_shards)
+  floor(((id + epoch_num) % num_shards + 1) * dataset_size / num_shards) -
+      floor(((id + epoch_num) % num_shards) * dataset_size / num_shards)
 
 When the second formula is used, providing a size value once at the beginning of the training works
 only when the ``stick_to_shard`` reader option is enabled and prevents DALI from rotating shards.

docs/auto_aug/auto_augment.rst

@@ -60,7 +60,8 @@ The best way is to wrap your policy creation into a function::
 
    def my_custom_policy() -> Policy:
         """
-        Creates a simple AutoAugment policy with 3 sub-policies using custom magnitude ranges.
+        Creates a simple AutoAugment policy with 3 sub-policies using custom
+        magnitude ranges.
         """
 
         shear_x = augmentations.shear_x.augmentation((0, 0.5), True)

docs/autodoc_submodules.py

@@ -45,7 +45,8 @@
 exclude_fn_members = {}
 
 installation_page_url = (
-    "https://docs.nvidia.com/deeplearning/dali/user-guide/docs/installation.html"
+    "https://docs.nvidia.com/deeplearning/dali/user-guide/"
+    "docs/installation.html"
 )
 
 mod_aditional_doc = {
@@ -87,11 +88,15 @@ def get_functions(module):
     or hidden members. No nested modules would be reported."""
     result = []
     # Take all public members of given module
-    public_members = list(filter(lambda x: not str(x).startswith("_"), dir(module)))
+    public_members = list(
+        filter(lambda x: not str(x).startswith("_"), dir(module))
+    )
     for member_name in public_members:
         member = getattr(module, member_name)
         # Just user-defined functions
-        if inspect.isfunction(member) and not member.__module__.endswith("hidden"):
+        if inspect.isfunction(member) and not member.__module__.endswith(
+            "hidden"
+        ):
             result.append(member_name)
     return result
 
@@ -157,7 +162,9 @@ def single_module_file(module, funs_in_module, references):
     result += "\n"
 
     result += f"The following table lists all operations available in ``{module}`` module:\n"
-    result += operations_table.operations_table_str(get_schema_names(module, funs_in_module))
+    result += operations_table.operations_table_str(
+        get_schema_names(module, funs_in_module)
+    )
     result += "\n\n"
 
     result += ".. toctree::\n   :hidden:\n\n"
@@ -185,15 +192,22 @@ def fn_autodoc(out_filename, generated_path, references):
         # the rest is within the same directory, so there is no need for that
         all_modules_str += f"   {generated_path / module}\n"
 
-        single_module_str = single_module_file(module, funs_in_module, references)
+        single_module_str = single_module_file(
+            module, funs_in_module, references
+        )
         with open(generated_path / (module + ".rst"), "w") as module_file:
             module_file.write(single_module_str)
 
         for fun in funs_in_module:
             full_name = f"{module}.{fun}"
-            if module in exclude_fn_members and fun in exclude_fn_members[module]:
+            if (
+                module in exclude_fn_members
+                and fun in exclude_fn_members[module]
+            ):
                 continue
-            with open(generated_path / (full_name + ".rst"), "w") as function_file:
+            with open(
+                generated_path / (full_name + ".rst"), "w"
+            ) as function_file:
                 single_file_str = single_fun_file(full_name, references)
                 function_file.write(single_file_str)