processing_sarashina2_vision.py

# coding=utf-8
# Copyright 2025 the SB Intuitions.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Processor class for Srashina2Vision.
"""

from copy import deepcopy
from typing import List, Optional, Union

import numpy as np
import torch
import torch.nn.functional as F
from PIL import Image
from transformers import (
    AutoImageProcessor,
    PreTrainedTokenizer,
    Qwen2VLImageProcessor,
    StoppingCriteria,
    StoppingCriteriaList,
)
from transformers.feature_extraction_utils import BatchFeature
from transformers.image_transforms import (
    convert_to_rgb,
    to_channel_dimension_format,
)
from transformers.image_utils import (
    ChannelDimension,
    ImageInput,
    VideoInput,
    get_image_size,
    infer_channel_dimension_format,
    is_scaled_image,
    make_list_of_images,
    to_numpy_array,
)
from transformers.models.qwen2_vl.image_processing_qwen2_vl import smart_resize
from transformers.processing_utils import ProcessingKwargs, ProcessorMixin, Unpack
from transformers.tokenization_utils_base import PreTokenizedInput, TextInput
from transformers.utils import logging

logger = logging.get_logger(__name__)


class GenerationStopper(StoppingCriteria):
    def __init__(
        self,
        stop_str_list: list[str],
        tokenizer: PreTrainedTokenizer,
        decode_suffix_length: int = 5,
    ):
        self.stop_str_list = stop_str_list
        self.tokenizer = deepcopy(tokenizer)
        self.decode_suffix_length = decode_suffix_length
        self.input_ids_end = None

    def __repr__(self):
        return f"Stopping words: {self.stop_str_list}"

    def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
        if self.input_ids_end is None:
            length = input_ids.shape[1]
            self.input_ids_end = length - 1 if (length - 1) > 0 else 0
        decode_ids = input_ids[0][self.input_ids_end :][-self.decode_suffix_length :]
        if len(decode_ids) == 0:
            decoded = ""
        else:
            decoded = self.tokenizer.decode(decode_ids)

        for stop_str in self.stop_str_list:
            if stop_str in decoded:
                self.input_ids_end = None
                return True
        return False

    @property
    def criteria(self):
        return StoppingCriteriaList([self])

    def format(self, sentence: str):
        for w in self.stop_str_list:
            if w in sentence[-len(w) :]:
                sentence = sentence[: -len(w)]
        return sentence


class Sarashina2VisionImageProcessor(Qwen2VLImageProcessor):
    def _preprocess(
        self,
        images: Union[ImageInput, VideoInput],
        do_resize: bool = None,
        resample: Image.Resampling = None,
        do_rescale: bool = None,
        rescale_factor: float = None,
        do_normalize: bool = None,
        image_mean: Optional[Union[float, List[float]]] = None,
        image_std: Optional[Union[float, List[float]]] = None,
        do_convert_rgb: bool = None,
        data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
        input_data_format: Optional[Union[str, ChannelDimension]] = None,
    ):
        """
        Preprocess an image or batch of images. Copy of the `preprocess` method from `Qwen2VLImageProcessor`.

        Args:
            images (`ImageInput`):
                Image or batch of images to preprocess. Expects pixel values ranging from 0 to 255. If pixel values range from 0 to 1, set `do_rescale=False`.
            vision_info (`List[Dict]`, *optional*):
                Optional list of dictionaries containing additional information about vision inputs.
            do_resize (`bool`, *optional*, defaults to `self.do_resize`):
                Whether to resize the image.
            resample (`PILImageResampling`, *optional*, defaults to `self.resample`):
                Resampling filter to use if resizing the image. This can be one of the `PILImageResampling` enums.
            do_rescale (`bool`, *optional*, defaults to `self.do_rescale`):
                Whether to rescale the image.
            rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`):
                Scale factor to use if rescaling the image.
            do_normalize (`bool`, *optional*, defaults to `self.do_normalize`):
                Whether to normalize the image.
            image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`):
                Mean to use if normalizing the image. Can be a float or a list of floats corresponding to the number of channels in the image.
            image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`):
                Standard deviation to use if normalizing the image. Can be a float or a list of floats corresponding to the number of channels in the image.
            do_convert_rgb (`bool`, *optional*, defaults to `self.do_convert_rgb`):
                Whether to convert the image to RGB.
            data_format (`ChannelDimension`, *optional*, defaults to `ChannelDimension.FIRST`):
                The channel dimension format for the output image. Can be one of:
                - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
                - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
                - Unset: Use the channel dimension format of the input image.
            input_data_format (`ChannelDimension` or `str`, *optional*):
                The channel dimension format for the input image. Can be one of:
                - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
                - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
                - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.   - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
        """
        images = make_list_of_images(images)

        if do_convert_rgb:
            images = [convert_to_rgb(image) for image in images]

        # All transformations expect numpy arrays.
        images = [to_numpy_array(image) for image in images]

        if do_rescale and is_scaled_image(images[0]):
            logger.warning_once(
                "It looks like you are trying to rescale already rescaled images. If the input"
                " images have pixel values between 0 and 1, set `do_rescale=False` to avoid rescaling them again."
            )
        if input_data_format is None:
            # We assume that all images have the same channel dimension format.
            input_data_format = infer_channel_dimension_format(images[0])

        height, width = get_image_size(images[0], channel_dim=input_data_format)
        resized_height, resized_width = height, width
        processed_images = []
        for image in images:
            if do_rescale:
                image = self.rescale(
                    image, scale=rescale_factor, input_data_format=input_data_format
                )

            if do_normalize:
                image = self.normalize(
                    image=image,
                    mean=image_mean,
                    std=image_std,
                    input_data_format=input_data_format,
                )

            image = to_channel_dimension_format(
                image, data_format, input_channel_dim=input_data_format
            )

            if do_resize:
                resized_height, resized_width = smart_resize(
                    height,
                    width,
                    factor=self.patch_size * self.merge_size,
                    min_pixels=self.min_pixels,
                    max_pixels=self.max_pixels,
                )
                image = (
                    F.interpolate(
                        torch.from_numpy(image).unsqueeze(0),
                        size=(resized_height, resized_width),
                        mode="bicubic",
                    )
                    .squeeze(0)
                    .numpy()
                )

            processed_images.append(image)

        patches = np.array(processed_images)
        if data_format == ChannelDimension.LAST:
            patches = patches.transpose(0, 3, 1, 2)
        if patches.shape[0] % self.temporal_patch_size != 0:
            repeats = np.repeat(patches[-1][np.newaxis], self.temporal_patch_size - 1, axis=0)
            patches = np.concatenate([patches, repeats], axis=0)
        channel = patches.shape[1]
        grid_t = patches.shape[0] // self.temporal_patch_size
        grid_h, grid_w = resized_height // self.patch_size, resized_width // self.patch_size
        patches = patches.reshape(
            grid_t,
            self.temporal_patch_size,
            channel,
            grid_h // self.merge_size,
            self.merge_size,
            self.patch_size,
            grid_w // self.merge_size,
            self.merge_size,
            self.patch_size,
        )
        patches = patches.transpose(0, 3, 6, 4, 7, 2, 1, 5, 8)
        flatten_patches = patches.reshape(
            grid_t * grid_h * grid_w,
            channel * self.temporal_patch_size * self.patch_size * self.patch_size,
        )

        return flatten_patches, (grid_t, grid_h, grid_w)


class Srashina2VisionProcessorKwargs(ProcessingKwargs, total=False):
    _defaults = {
        "text_kwargs": {
            "padding": False,
        },
    }


class Srashina2VisionProcessor(ProcessorMixin):
    r"""
    Constructs Srashina2Vision processor which wraps a Srashina2Vision image processor and a LLama tokenizer into a single processor.
    [`Srashina2VisionProcessor`] offers all the functionalities of [`Sarashina2VisionImageProcessor`] and [`LlamaTokenizerFast`]. See the
    [`~Srashina2VisionProcessor.__call__`] and [`~Srashina2VisionProcessor.decode`] for more information.
    Args:
        image_processor ([`Sarashina2VisionImageProcessor`], *optional*):
            The image processor is a required input.
        tokenizer ([`LlamaTokenizerFast`], *optional*):
            The tokenizer is a required input.
        chat_template (`str`, *optional*): A Jinja template which will be used to convert lists of messages
            in a chat into a tokenizable string.
    """

    attributes = ["image_processor", "tokenizer"]
    valid_kwargs = ["chat_template"]
    image_processor_class = "AutoImageProcessor"
    tokenizer_class = ("LlamaTokenizer", "LlamaTokenizerFast")

    def __init__(self, image_processor=None, tokenizer=None, chat_template=None, **kwargs):
        self.image_token = (
            "<|file|>" if not hasattr(tokenizer, "image_token") else tokenizer.image_token
        )
        self.stop_symbol = "\n###"
        super().__init__(image_processor, tokenizer, chat_template=chat_template)

    def __call__(
        self,
        images: ImageInput = None,
        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
        **kwargs: Unpack[Srashina2VisionProcessorKwargs],
    ) -> BatchFeature:
        """
        Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
        and `kwargs` arguments to LlamaTokenizerFast's [`~LlamaTokenizerFast.__call__`] if `text` is not `None` to encode
        the text. To prepare the vision inputs, this method forwards the `vision_infos` and `kwrags` arguments to
        Sarashina2VisionImageProcessor's [`~Sarashina2VisionImageProcessor.__call__`] if `vision_infos` is not `None`.

        Args:
            images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
                The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
                tensor. Both channels-first and channels-last formats are supported.
            text (`str`, `List[str]`, `List[List[str]]`):
                The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings
                (pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set
                `is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
            return_tensors (`str` or [`~utils.TensorType`], *optional*):
                If set, will return tensors of a particular framework. Acceptable values are:
                - `'tf'`: Return TensorFlow `tf.constant` objects.
                - `'pt'`: Return PyTorch `torch.Tensor` objects.
                - `'np'`: Return NumPy `np.ndarray` objects.
                - `'jax'`: Return JAX `jnp.ndarray` objects.

        Returns:
            [`BatchFeature`]: A [`BatchFeature`] with the following fields:

            - **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`.
            - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when
              `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not
              `None`).
            - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`.
            - **image_grid_thw** -- List of image 3D grid in LLM. Returned when `images` is not `None`.
        """
        output_kwargs = self._merge_kwargs(
            Srashina2VisionProcessorKwargs,
            tokenizer_init_kwargs=self.tokenizer.init_kwargs,
            **kwargs,
        )
        if images is not None:
            image_inputs = self.image_processor(
                images=images, videos=None, **output_kwargs["images_kwargs"]
            )
            image_grid_thw = image_inputs["image_grid_thw"]
        else:
            image_inputs = {}
            image_grid_thw = None

        if not isinstance(text, list):
            text = [text]

        if image_grid_thw is not None:
            merge_length = self.image_processor.merge_size**2
            index = 0
            for i in range(len(text)):
                while self.image_token in text[i]:
                    text[i] = text[i].replace(
                        self.image_token,
                        "<|placeholder|>" * (image_grid_thw[index].prod() // merge_length),
                        1,
                    )
                    index += 1
                text[i] = text[i].replace("<|placeholder|>", self.image_token)

        text_inputs = self.tokenizer(text, **output_kwargs["text_kwargs"])

        return BatchFeature(data={**text_inputs, **image_inputs})

    def batch_decode(self, *args, **kwargs):
        """
        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`].
        """
        return [
            output.replace(self.stop_symbol, "")
            for output in self.tokenizer.batch_decode(*args, **kwargs)
        ]

    def decode(self, *args, **kwargs):
        """
        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`].
        """
        return self.tokenizer.decode(*args, **kwargs).replace(self.stop_symbol, "")

    def post_process_image_text_to_text(self, generated_outputs):
        """
        Post-process the output of the model to decode the text.

        Args:
            generated_outputs (`torch.Tensor` or `np.ndarray`):
                The output of the model `generate` function. The output is expected to be a tensor of shape `(batch_size, sequence_length)`
                or `(sequence_length,)`.

        Returns:
            `List[str]`: The decoded text.
        """
        return self.tokenizer.batch_decode(
            generated_outputs, skip_special_tokens=True, clean_up_tokenization_spaces=False
        )

    @property
    def model_input_names(self):
        tokenizer_input_names = self.tokenizer.model_input_names
        image_processor_input_names = self.image_processor.model_input_names
        return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))

    def get_stopping_criteria(self, stop_symbols: List[str]):
        stopping_criteria = GenerationStopper(stop_str_list=stop_symbols, tokenizer=self.tokenizer)
        return stopping_criteria.criteria


Srashina2VisionProcessor.register_for_auto_class("AutoProcessor")
AutoImageProcessor.register("Sarashina2VisionImageProcessor", Sarashina2VisionImageProcessor)