clarifai 9.7.1__py3-none-any.whl → 9.7.2__py3-none-any.whl

clarifai_utils/data_upload/examples/image_segmentation/coco/dataset.py

@@ -1,107 +0,0 @@
-#! COCO 2017 Image Segmentation dataset
-
-import gc
-import os
-from functools import reduce
-
-import cv2
-import numpy as np
-from pycocotools import mask as maskUtils
-from pycocotools.coco import COCO
-
-from clarifai.data_upload.datasets.features import VisualSegmentationFeatures
-
-
-class COCOSegmentationDataset:
-  """COCO 2017 Image Segmentation Dataset.
-  url: https://cocodataset.org/#download
-  """
-
-  def __init__(self, split: str = "train"):
-    """
-    Inititalize dataset params.
-    Args:
-      split: "train" or "test"
-    """
-    self.split = split
-    self.image_dir = {"train": os.path.join(os.path.dirname(__file__), "images")}
-    self.annotations_file = {
-        "train":
-            os.path.join(os.path.dirname(__file__), "annotations/instances_val2017_subset.json")
-    }
-
-  def dataloader(self):
-    """
-    Transform COCO 2017 segmentation dataset into clarifai proto compatible
-    format to uplaod
-    Returns:
-      VisualSegmentationFeatures type generator.
-    """
-    coco = COCO(self.annotations_file[self.split])
-    categories = coco.loadCats(coco.getCatIds())
-    cat_id_map = {category["id"]: category["name"] for category in categories}
-    cat_img_ids = {}
-    for cat_id in list(cat_id_map.keys()):
-      cat_img_ids[cat_id] = coco.getImgIds(catIds=[cat_id])
-
-    img_ids = []
-    for i in list(cat_img_ids.values()):
-      img_ids.extend(i)
-
-    # Get the image information for the specified image IDs
-    image_info = coco.loadImgs(img_ids)
-    # Extract the file names from the image information
-    image_filenames = {img_id: info['file_name'] for info, img_id in zip(image_info, img_ids)}
-    #get annotations for each image id
-    for _id in set(img_ids):
-      annots = []  # polygons
-      class_names = []
-      labels = [i for i in list(filter(lambda x: _id in cat_img_ids[x], cat_img_ids))]
-      image_path = os.path.join(self.image_dir[self.split], image_filenames[_id])
-
-      image_height, image_width = cv2.imread(image_path).shape[:2]
-      for cat_id in labels:
-        annot_ids = coco.getAnnIds(imgIds=_id, catIds=[cat_id])
-
-        if len(annot_ids) > 0:
-          img_annotations = coco.loadAnns(annot_ids)
-          for ann in img_annotations:
-            # get polygons
-            if type(ann['segmentation']) == list:
-              for seg in ann['segmentation']:
-                poly = np.array(seg).reshape((int(len(seg) / 2), 2))
-                poly[:, 0], poly[:, 1] = poly[:, 0] / image_width, poly[:, 1] / image_height
-                annots.append(poly.tolist())  #[[x=col, y=row],...]
-                class_names.append(cat_id_map[cat_id])
-            else:  # seg: {"counts":[...]}
-              if type(ann['segmentation']['counts']) == list:
-                rle = maskUtils.frPyObjects([ann['segmentation']], image_height, image_width)
-              else:
-                rle = ann['segmentation']
-              mask = maskUtils.decode(rle)  #binary mask
-              #convert mask to polygons and add to annots
-              contours, _ = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
-              polygons = []
-              for cont in contours:
-                if cont.size >= 6:
-                  polygons.append(cont.astype(float).flatten().tolist())
-              # store polygons in (x,y) pairs
-              polygons_flattened = reduce(lambda x, y: x + y, polygons)
-              del polygons
-              del contours
-              del mask
-              gc.collect()
-
-              polygons = np.array(polygons_flattened).reshape((int(len(polygons_flattened) / 2),
-                                                               2))
-              polygons[:, 0] = polygons[:, 0] / image_width
-              polygons[:, 1] = polygons[:, 1] / image_height
-
-              annots.append(polygons.tolist())  #[[x=col, y=row],...,[x=col, y=row]]
-              class_names.append(cat_id_map[cat_id])
-        else:  # if no annotations for given image_id-cat_id pair
-          continue
-      assert len(class_names) == len(annots), f"Num classes must match num annotations\
-      for a single image. Found {len(class_names)} classes and {len(annots)} polygons."
-
-      yield VisualSegmentationFeatures(image_path, class_names, annots, id=_id)

clarifai_utils/data_upload/examples/text_classification/imdb_dataset/dataset.py

@@ -1,40 +0,0 @@
-#! IMDB 50k Movie Reviews dataset
-
-import csv
-import os
-
-from clarifai.data_upload.datasets.features import TextFeatures
-
-
-class IMDBMovieReviewsDataset:
-  """IMDB 50K Movie Reviews Dataset."""
-
-  def __init__(self, split: str = "train"):
-    """
-    Initialize dataset params.
-    Args:
-      data_dir: the local dataset directory.
-      split: "train" or "test"
-    """
-    self.split = split
-    self.data_dirs = {
-        "train": os.path.join(os.path.dirname(__file__), "train.csv"),
-        "test": os.path.join(os.path.dirname(__file__), "test.csv")
-    }
-
-  def dataloader(self):
-    """
-    Transform text data into clarifai proto compatible
-    format for upload.
-    Returns:
-      TextFeatures type generator.
-    """
-    ## Your preprocessing code here
-    with open(self.data_dirs[self.split]) as _file:
-      reader = csv.reader(_file)
-      next(reader, None)  # skip header
-      for review in reader:
-        yield TextFeatures(
-            text=review[0],  # text,
-            labels=review[1],  # sentiment,
-            id=None)

clarifai_utils/data_upload/examples.py

@@ -1,17 +0,0 @@
-#! Execute dataset upload using the `from_module` upload feature
-
-from clarifai.data_upload.upload import UploadConfig
-
-text_upload_obj = UploadConfig(
-    user_id="",
-    app_id="",
-    pat="",
-    dataset_id="",
-    task="visual_clf",
-    from_module="./examples/image_classification/cifar10",
-    split="train")
-## change the task and from_module arguments in UploadConfig() to upload
-## example food-101 dataset
-
-if __name__ == "__main__":
-  text_upload_obj.upload_to_clarifai()

clarifai_utils/data_upload/upload.py

@@ -1,356 +0,0 @@
-#! Clarifai data upload
-
-import importlib
-import inspect
-import os
-import sys
-import time
-import uuid
-from concurrent.futures import ThreadPoolExecutor, as_completed
-from multiprocessing import cpu_count
-from typing import Iterator, List, Optional, Tuple, Union
-
-from clarifai_grpc.grpc.api import resources_pb2, service_pb2, service_pb2_grpc
-from clarifai_grpc.grpc.api.status import status_code_pb2, status_pb2
-from google.protobuf.json_format import MessageToDict
-from tqdm import tqdm
-from clarifai.auth.helper import ClarifaiAuthHelper
-from clarifai.client import create_stub
-from clarifai.data_upload.datasets.base import Chunker
-from clarifai.data_upload.datasets.image import (VisualClassificationDataset,
-                                                 VisualDetectionDataset, VisualSegmentationDataset)
-from clarifai.data_upload.datasets.text import TextClassificationDataset
-
-
-def load_dataset(module_dir: Union[str, os.PathLike], split: str) -> Iterator:
-  """
-  Validate and import dataset module data generator.
-  Args:
-    `module_dir`: relative path to the module directory
-    The directory must contain a `dataset.py` script and the data itself.
-    `split`: "train" or "val"/"test" dataset split
-  Module Directory Structure:
-  ---------------------------
-      <folder_name>/
-      ├──__init__.py
-      ├──<Your local dir dataset>/
-      └──dataset.py
-  dataset.py must implement a class named following the convention,
-  <dataset_name>Dataset and this class must have a dataloader()
-  generator method
-  """
-  sys.path.append(str(module_dir))
-
-  if not os.path.exists(os.path.join(module_dir, "__init__.py")):
-    with open(os.path.join(module_dir, "__init__.py"), "w"):
-      pass
-
-  import dataset  # dataset module
-
-  # get main module class
-  main_module_cls = None
-  for name, obj in dataset.__dict__.items():
-    if inspect.isclass(obj) and "Dataset" in name:
-      main_module_cls = obj
-    else:
-      continue
-
-  return main_module_cls(split).dataloader()
-
-
-def load_zoo_dataset(name: str, split: str) -> Iterator:
-  """
-  Get dataset generator object from dataset zoo.
-  Args:
-    `name`: dataset module name in datasets/zoo/.
-    `split`: "train" or "val"/"test" dataset split
-  Returns:
-    Data generator object
-  """
-  zoo_dataset = importlib.import_module(f"datasets.zoo.{name}")
-  # get main module class
-  main_module_cls = None
-  for name, obj in zoo_dataset.__dict__.items():
-    if inspect.isclass(obj) and "Dataset" in name:
-      main_module_cls = obj
-    else:
-      continue
-
-  return main_module_cls(split).dataloader()
-
-
-class UploadConfig:
-
-  def __init__(
-      self,
-      user_id: str,
-      app_id: str,
-      pat: str,
-      dataset_id: str,
-      task: str,
-      from_module: Optional[Union[str, os.PathLike]] = None,
-      from_zoo: Optional[str] = None,  # load dataset from zoo
-      split: str = "train",  # train or test/val
-      chunk_size: int = 128,
-      base_url: str = None):
-    """
-    Initialize upload configs.
-    Args:
-      `user_id`: Clarifai user id.
-      `app_id`: Clarifai app id.
-      `pat`: Clarifai PAT(Personal Access Token).
-      `dataset_id`: Clarifai dataset id (where data is to be uploaded).
-      `task`: either of `visual_clf`, `visual_detection`, `visual_segmentation` or `text_clf`.
-      `from_module`: Path to dataset module directory.
-        Should be left as None if `from_zoo` is to be used.
-      `from_zoo`: Name of dataset to upload from the zoo.
-        The name must match the dataset module name excluding the file extension.
-        Should be left as None if `from_module` is to be used.
-      `split`: Dataset split to upload. Either of train or test/val
-      `chunk_size`: size of chunks for parallel data upload.
-    """
-    self.USER_ID = user_id
-    self.APP_ID = app_id
-    self.PAT = pat
-    self.dataset_id = dataset_id
-    self.task = task
-    self.module_dir = from_module
-    self.zoo_dataset = from_zoo
-    self.split = split
-    self.chunk_size = min(128, chunk_size)  # limit max protos in a req
-    self.num_workers: int = min(10, cpu_count())  #15 req/sec rate limit
-    self.annot_num_workers = 4
-    self.max_retires = 10
-    self.__base: str = ""
-    if not base_url:
-      self.__base = "https://api.clarifai.com"
-    else:
-      self.__base = base_url
-
-    auth_helper = ClarifaiAuthHelper(
-        user_id=self.USER_ID, app_id=self.APP_ID, pat=self.PAT, base=self.__base)
-    self.STUB: service_pb2_grpc.V2Stub = create_stub(auth_helper)
-    self.user_app_id = auth_helper.get_user_app_id_proto()
-
-  def _upload_inputs(self, batch_input: List[resources_pb2.Input]) -> str:
-    """
-    Upload inputs to clarifai platform dataset.
-    Args:
-      batch_input: input batch protos
-    Returns:
-      input_job_id: Upload Input Job ID
-    """
-    input_job_id = uuid.uuid4().hex  # generate a unique id for this job
-    response = self.STUB.PostInputs(
-        service_pb2.PostInputsRequest(
-            user_app_id=self.user_app_id, inputs=batch_input, inputs_add_job_id=input_job_id),)
-    if response.status.code != status_code_pb2.SUCCESS:
-      try:
-        print(f"Post inputs failed, status: {response.inputs[0].status.details}")
-      except:
-        print(f"Post inputs failed, status: {response.status.details}")
-
-    return input_job_id
-
-  def _upload_annotations(self, batch_annot: List[resources_pb2.Annotation]
-                         ) -> Union[List[resources_pb2.Annotation], List[None]]:
-    """
-    Upload image annotations to clarifai detection dataset
-    Args:
-      batch_annot: annot batch protos
-    Returns:
-      retry_upload: failed annot upload
-    """
-    retry_upload = []  # those that fail to upload are stored for retries
-    response = self.STUB.PostAnnotations(
-        service_pb2.PostAnnotationsRequest(user_app_id=self.user_app_id, annotations=batch_annot),)
-
-    if response.status.code != status_code_pb2.SUCCESS:
-      try:
-        print(f"Post annotations failed, status: {response.annotations[0].status.details}")
-      except:
-        print(f"Post annotations failed, status: {response.status.details}")
-      finally:
-        retry_upload.extend(batch_annot)
-
-    return retry_upload
-
-  def _concurrent_annot_upload(self, annots: List[List[resources_pb2.Annotation]]
-                              ) -> Union[List[resources_pb2.Annotation], List[None]]:
-    """
-    Uploads annotations concurrently.
-    Args:
-      annots: annot protos
-    Returns:
-      retry_annot_upload: All failed annot protos during upload
-    """
-    annot_threads = []
-    retry_annot_upload = []
-
-    with ThreadPoolExecutor(max_workers=self.annot_num_workers) as executor:  # limit annot workers
-      annot_threads = [
-          executor.submit(self._upload_annotations, inp_batch) for inp_batch in annots
-      ]
-
-      for job in as_completed(annot_threads):
-        result = job.result()
-        if result:
-          retry_annot_upload.extend(result)
-
-    return retry_annot_upload
-
-  def _backoff_iterator(self) -> None:
-    """
-    Return iterator for exponential backoff intervals.
-    """
-    yield 0.1
-    for i in range(5, 11):
-      yield 0.01 * (2**i)
-    while True:
-      yield 0.01 * (2**10)  #10 sec
-
-  def _wait_for_inputs(self, input_job_id: str) -> bool:
-    """
-    Wait for inputs to be processed. Cancel Job if timeout > 30 minutes.
-    Args:
-      input_job_id: Upload Input Job ID
-    Returns:
-      True if inputs are processed, False otherwise
-    """
-    backoff_iterator = self._backoff_iterator()
-    max_retries = self.max_retires
-    start_time = time.time()
-    while True:
-      response = self.STUB.GetInputsAddJob(
-          service_pb2.GetInputsAddJobRequest(user_app_id=self.user_app_id, id=input_job_id),)
-
-      if time.time() - start_time > 60 * 30 or max_retries == 0:  # 30 minutes timeout
-        self.STUB.CancelInputsAddJob(
-            service_pb2.CancelInputsAddJobRequest(user_app_id=self.user_app_id, id=input_job_id),
-        )  #Cancel Job
-        return False
-      if response.status.code != status_code_pb2.SUCCESS:
-        max_retries -= 1
-        print(f"Get input job failed, status: {response.status.details}\n")
-        continue
-      if response.inputs_add_job.progress.in_progress_count == 0 and response.inputs_add_job.progress.pending_count == 0:
-        return True
-      else:
-        time.sleep(next(backoff_iterator))
-
-  def _delete_failed_inputs(self, input_ids: List[str]) -> Tuple[List[str], List[str]]:
-    """
-    Delete failed input ids from clarifai platform dataset.
-    Args:
-      input_ids: batch input ids
-    Returns:
-      success_inputs: upload success input ids
-      failed_inputs: upload failed input ids
-    """
-    success_status = status_pb2.Status(code=status_code_pb2.INPUT_DOWNLOAD_SUCCESS)
-    response = self.STUB.ListInputs(
-        service_pb2.ListInputsRequest(
-            ids=input_ids,
-            per_page=len(input_ids),
-            user_app_id=self.user_app_id,
-            status=success_status),)
-    response_dict = MessageToDict(response)
-    success_inputs = response_dict.get('inputs', [])
-
-    success_input_ids = [input.get('id') for input in success_inputs]
-    failed_input_ids = list(set(input_ids) - set(success_input_ids))
-    #delete failed inputs
-    self.STUB.DeleteInputs(
-        service_pb2.DeleteInputsRequest(user_app_id=self.user_app_id, ids=failed_input_ids),)
-
-    return success_input_ids, failed_input_ids
-
-  def _upload_inputs_annotations(
-      self, batch_input_ids: List[str]) -> Tuple[List[str], List[resources_pb2.Annotation]]:
-    """
-    Uploads batch of inputs and annotations concurrently to clarifai platform dataset.
-    Args:
-      batch_input_ids: batch input ids
-    Returns:
-      failed_input_ids: failed input ids
-      retry_annot_protos: failed annot protos
-    """
-    input_protos, _ = self.dataset_obj.get_protos(batch_input_ids)
-    input_job_id = self._upload_inputs(input_protos)
-    retry_annot_protos = []
-
-    self._wait_for_inputs(input_job_id)
-    success_input_ids, failed_input_ids = self._delete_failed_inputs(batch_input_ids)
-
-    if self.task in ["visual_detection", "visual_segmentation"]:
-      _, annotation_protos = self.dataset_obj.get_protos(success_input_ids)
-      chunked_annotation_protos = Chunker(annotation_protos, self.chunk_size).chunk()
-      retry_annot_protos.extend(self._concurrent_annot_upload(chunked_annotation_protos))
-
-    return failed_input_ids, retry_annot_protos
-
-  def _retry_uploads(self, failed_input_ids: List[str],
-                     retry_annot_protos: List[resources_pb2.Annotation]) -> None:
-    """
-    Retry failed uploads.
-    Args:
-      failed_input_ids: failed input ids
-      retry_annot_protos: failed annot protos
-    """
-    if failed_input_ids:
-      self._upload_inputs_annotations(failed_input_ids)
-    if retry_annot_protos:
-      chunked_annotation_protos = Chunker(retry_annot_protos, self.chunk_size).chunk()
-      _ = self._concurrent_annot_upload(chunked_annotation_protos)
-
-  def _data_upload(self, input_ids: List[str]) -> None:
-    """
-    Uploads inputs and annotations to clarifai platform dataset.
-    Args:
-      input_ids: input ids
-    """
-    chunk_input_ids = Chunker(input_ids, self.chunk_size).chunk()
-    with ThreadPoolExecutor(max_workers=self.num_workers) as executor:
-      with tqdm(total=len(chunk_input_ids), desc='Uploading Dataset') as progress:
-        # Submit all jobs to the executor and store the returned futures
-        futures = [
-            executor.submit(self._upload_inputs_annotations, batch_input_ids)
-            for batch_input_ids in chunk_input_ids
-        ]
-
-        for job in as_completed(futures):
-          retry_input_proto, retry_annot_protos = job.result()
-          self._retry_uploads(retry_input_proto, retry_annot_protos)
-          progress.update()
-
-  def upload_to_clarifai(self):
-    """
-    Execute data upload.
-    """
-    datagen_object = None
-    if self.module_dir is None and self.zoo_dataset is None:
-      raise Exception("One of `from_module` and `from_zoo` must be \
-      specified. Both can't be None or defined at the same time.")
-    elif self.module_dir is not None and self.zoo_dataset is not None:
-      raise Exception("Use either of `from_module` or `from_zoo` \
-      but NOT both.")
-    elif self.module_dir is not None:
-      datagen_object = load_dataset(self.module_dir, self.split)
-    else:
-      datagen_object = load_zoo_dataset(self.zoo_dataset, self.split)
-
-    if self.task == "text_clf":
-      self.dataset_obj = TextClassificationDataset(datagen_object, self.dataset_id, self.split)
-      self._data_upload(self.dataset_obj.input_ids)
-
-    elif self.task == "visual_detection":
-      self.dataset_obj = VisualDetectionDataset(datagen_object, self.dataset_id, self.split)
-      self._data_upload(self.dataset_obj.input_ids)  # TODO: get_img_ids or get_input_ids
-
-    elif self.task == "visual_segmentation":
-      self.dataset_obj = VisualSegmentationDataset(datagen_object, self.dataset_id, self.split)
-      self._data_upload(self.dataset_obj.input_ids)
-
-    else:  # visual-classification & visual-captioning
-      self.dataset_obj = VisualClassificationDataset(datagen_object, self.dataset_id, self.split)
-      self._data_upload(self.dataset_obj.input_ids)