processing_omchat.py

from typing import List, Optional, Union

from transformers import PreTrainedTokenizer
from typing import List, Tuple

from transformers.feature_extraction_utils import BatchFeature
from transformers.image_utils import ImageInput
from transformers.processing_utils import ProcessorMixin
from transformers.tokenization_utils_base import PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
from transformers.utils import TensorType
import torch

def tokenizer_image_token(prompt, tokenizer, image_token_index=-200, return_tensors=None):
    if "<image_0>" in prompt:
        image_token_pattern = re.compile(r"<image_(\d+)>")
        prompt_chunks = re.split(r'<image_[0-9]+>',prompt)
        # Identify all the image tags
        image_tags = image_token_pattern.findall(prompt)

        input_ids = []
        for i, chunk in enumerate(prompt_chunks):
            input_ids.extend(tokenizer(chunk).input_ids)
            if i < len(image_tags):
                #input_ids.append(-100 * (int(image_tags[i]) + 3))
                input_ids.append(-200)
    else:
        prompt_chunks = [tokenizer(chunk).input_ids for chunk in prompt.split('<image>')]
        def insert_separator(X, sep):
            return [ele for sublist in zip(X, [sep]*len(X)) for ele in sublist][:-1]

        input_ids = []
        offset = 0
        if len(prompt_chunks) > 0 and len(prompt_chunks[0]) > 0 and prompt_chunks[0][0] == tokenizer.bos_token_id:
            offset = 1
            input_ids.append(prompt_chunks[0][0])

        for x in insert_separator(prompt_chunks, [image_token_index] * (offset + 1)):
            input_ids.extend(x[offset:])
    # Convert to tensor if required
    if return_tensors is not None:
        if return_tensors == 'pt':
            return torch.tensor(input_ids, dtype=torch.long)
        else:
            raise ValueError(f'Unsupported tensor type: {return_tensors}')

    return input_ids


def make_context(
    tokenizer: PreTrainedTokenizer,
    query: str,
    history: List[Tuple[str, str]] = None,
    system: str = "",
    max_window_size: int = 6144,
    chat_format: str = "chatml",
):
    if history is None:
        history = []

    if chat_format == "chatml":
        im_start, im_end = "<|im_start|>", "<|im_end|>"
        im_start_tokens = [151644]
        im_end_tokens = [151645]
        nl_tokens = tokenizer.encode("\n")

        def _tokenize_str(role, content):
            if "<image>" in content:
                return f"{role}\n{content}", tokenizer.encode(
                    role
                ) + nl_tokens + tokenizer_image_token(
                    content, tokenizer, -200
                )
            else:
                return f"{role}\n{content}", tokenizer.encode(
                    role
                ) + nl_tokens + tokenizer.encode(content)

        def _tokenize_str2(role, content):
            return f"{role}\n{content}", tokenizer.encode(
                role,
            ) + nl_tokens + tokenizer.encode(content)

        system_text, system_tokens_part = _tokenize_str("system", system)
        system_tokens = im_start_tokens + system_tokens_part + im_end_tokens

        raw_text = ""
        context_tokens = []

        for turn_query, turn_response in reversed(history):
            query_text, query_tokens_part = _tokenize_str("user", turn_query)
            query_tokens = im_start_tokens + query_tokens_part + im_end_tokens
            response_text, response_tokens_part = _tokenize_str(
                "assistant", turn_response
            )
            response_tokens = im_start_tokens + response_tokens_part + im_end_tokens

            next_context_tokens = nl_tokens + query_tokens + nl_tokens + response_tokens
            prev_chat = (
                f"\n{im_start}{query_text}{im_end}\n{im_start}{response_text}{im_end}"
            )

            current_context_size = (
                len(system_tokens) + len(next_context_tokens) + len(context_tokens)
            )
            if current_context_size < max_window_size:
                context_tokens = next_context_tokens + context_tokens
                raw_text = prev_chat + raw_text
            else:
                break

        context_tokens = system_tokens + context_tokens
        raw_text = f"{im_start}{system_text}{im_end}" + raw_text
        context_tokens += (
            nl_tokens
            + im_start_tokens
            + _tokenize_str("user", query)[1]
            + im_end_tokens
            + nl_tokens
            + im_start_tokens
            + tokenizer.encode("assistant")
            + nl_tokens
        )
        raw_text += f"\n{im_start}user\n{query}{im_end}\n{im_start}assistant\n"

    elif chat_format == "raw":
        raw_text = query
        context_tokens = tokenizer.encode(raw_text)
    else:
        raise NotImplementedError(f"Unknown chat format {chat_format!r}")

    return raw_text, context_tokens

def split_tensor(A, B):
    split_tensors = []
    start_idx = 0

    for i, size in enumerate(B.tolist()):
        split_tensor = A[i, :size, :, :, :]
        split_tensors.append(split_tensor)  # Take the first element from the batch dimension

    return split_tensors

class OmChatProcessor(ProcessorMixin):
    r"""
    Constructs a OmChat processor which wraps a OmChat image processor and a LLaMa tokenizer into a single processor.

    [`OmChatProcessor`] offers all the functionalities of [`OmChatImageProcessor`] and [`LlamaTokenizerFast`]. See the
    [`~OmChatProcessor.__call__`] and [`~OmChatProcessor.decode`] for more information.

    Args:
        image_processor ([`OmChatImageProcessor`], *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 = "AutoTokenizer"

    def __init__(self, image_processor=None, tokenizer=None,  **kwargs):
        super().__init__(image_processor, tokenizer)

    def __call__(
        self,
        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]],
        system_prompt: str = "You are a helpful assistant.",
        images: ImageInput = None,
        padding: Union[bool, str, PaddingStrategy] = False,
        truncation: Union[bool, str, TruncationStrategy] = None,
        max_length: Optional[int] = None,
        return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,
    ) -> 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 image(s), this method forwards the `images` and `kwrags` arguments to
        OmChatImageProcessor's [`~OmChatImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring
        of the above two methods for more information.

        Args:
            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).
            system_prompt ('str'):
                the initial system prompt (i.e., You are a helpful assistant.)
            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.
            padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `False`):
                Select a strategy to pad the returned sequences (according to the model's padding side and padding
                index) among:
                - `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
                  sequence if provided).
                - `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum
                  acceptable input length for the model if that argument is not provided.
                - `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different
                  lengths).
            max_length (`int`, *optional*):
                Maximum length of the returned list and optionally padding length (see above).
            truncation (`bool`, *optional*):
                Activates truncation to cut input sequences longer than `max_length` to `max_length`.
            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`.
        """
        #system_prompt="A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions."
        if images is not None:
            image_inputs = self.image_processor(images, return_tensors=return_tensors)
            new_images = []
            new_texts = []
            img = image_inputs["pixel_values"]
            num_patches = image_inputs["num_patches"]
            img = split_tensor(img, num_patches)
            if len(img) == 1:
                n = num_patches.tolist()[0]
                inp, context_tokens = make_context(
                    self.tokenizer,
                    "<image>\n"+"\n".join(["patch:<image>"]*(n-1)) +"\n"+ text.replace("<image>", ""),
                    None,
                    system_prompt,
                )
            
            else:
                texts = text.split("<image>")
                final =texts[0]
                for i, n in enumerate(num_patches.tolist()):
                    final+= "\n<image>\n"+"\n".join(["patch:<image>"]*(n-1))+"\n"
                    if i+1 < len(texts):
                        final += texts[i+1]
                inp, context_tokens = make_context(self.tokenizer, final, None, system_prompt)
            text_inputs = {"input_ids": torch.tensor([context_tokens])}
            image_inputs = {"images":torch.cat(img, dim=0)}
            return BatchFeature(data={**text_inputs, **image_inputs})
        else:
            image_inputs = {"images":None}
            inp, context_tokens = make_context(
                    self.tokenizer,
                    text.replace("<image>", "").strip(),
                    None,
                    "You are a helpful assistant.",
                )
            text_inputs = {"input_ids": torch.tensor([context_tokens])}

            return BatchFeature(data={**text_inputs})

    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.batch_decode with CLIP->Llama
    def batch_decode(self, *args, **kwargs):
        """
        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please
        refer to the docstring of this method for more information.
        """
        return self.tokenizer.batch_decode(*args, **kwargs)

    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.decode with CLIP->Llama
    def decode(self, *args, **kwargs):
        """
        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to
        the docstring of this method for more information.
        """
        return self.tokenizer.decode(*args, **kwargs)

    @property
    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.model_input_names
    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))