o ϯi @sddlZddlZddlZddlZddlmZddlmZmZm Z m Z m Z m Z ddl Z ddlZ ddlmZddlmZddlmZddlmZGd d d e jjjZ dd e ed ede e ee effddZdS)N) defaultdict)AnyCallableListOptionalTupleType) g_pathmgr) ClipSampler) FrameVideo)MultiProcessSamplerc@seZdZdZdZejjjdddfde de de ejjj de eegefd e d e ed dfd d Zedeed ed eefddZed ejjj fddZd efddZddZdS)CharadesaV Action recognition video dataset for `Charades `_ stored as image frames. This dataset handles the parsing of frames, loading and clip sampling for the videos. All io is done through :code:`iopath.common.file_io.PathManager`, enabling non-local storage uri's to be used. N data_path clip_sampler video_sampler transformvideo_path_prefixframes_per_clipreturncCsvtjd||_||_t||d\|_|_|_||j|_ d|_ |dur-t j t j|dnd|_d|_d|_d|_dS)a Args: data_path (str): Path to the data file. This file must be a space separated csv with the format: (original_vido_id video_id frame_id path_labels) clip_sampler (ClipSampler): Defines how clips should be sampled from each video. See the clip sampling documentation for more information. video_sampler (Type[torch.utils.data.Sampler]): Sampler for the internal video container. This defines the order videos are decoded and, if necessary, the distributed split. transform (Optional[Callable]): This callable is evaluated on the clip output before the clip is returned. It can be used for user defined preprocessing and augmentations on the clips. The clip output format is described in __next__(). video_path_prefix (str): prefix path to add to all paths from data_path. frames_per_clip (Optional[int]): The number of frames per clip to sample. z&PYTORCHVIDEO.dataset.Charades.__init__)prefixN)r)torch_C_log_api_usage_once _transform _clip_sampler_read_video_paths_and_labels_path_to_videos_labels _video_labels_video_sampler_video_sampler_iter functoolspartialr_sample_clip_frames _frame_filter _loaded_video _loaded_clip_next_clip_start_time)selfrrrrrrr-N/home/ubuntu/.local/lib/python3.10/site-packages/pytorchvideo/data/charades.py__init__ s(     zCharades.__init__ frame_indicescsBt}td|d|}t|d|d}fdd|DS)z Args: frame_indices (list): list of frame indices. frames_per+clip (int): The number of frames per clip to sample. Returns: (list): Outputs a subsampled list with num_samples frames. rr cg|]}|qSr-r-).0idxr0r-r. jz0Charades._sample_clip_frames..)lenrlinspaceclamplong)r0r num_framesindicesr-r4r.r'Zs zCharades._sample_clip_framescCs|jSN)r#r,r-r-r.rlszCharades.video_samplerc sjs ttj_jrj\}ntjj}t|}|f_ j |j i\}}}}}|dkrC| ||j _jdjd}} |_ |rYd_d_ fddtt| t| dD} || jt||d } jdur| } | S) a Retrieves the next clip based on the clip sampling strategy and video sampler. Returns: A dictionary with the following format. .. code-block:: text { 'video': , 'label': , 'video_label': 'video_index': , 'clip_index': , 'aug_index': , } rvideor0Nrcsg|] }j|qSr-)r!)r2ir, video_indexr-r.r5s z%Charades.__next__..r )r?label video_label video_namerB clip_index aug_index)r$iterr r#r)nextr r from_frame_pathsrr+durationget_clipr(r*rangeminmaxr"strr) r,r?path_to_video_frames clip_startclip_endrFrG is_last_clipframesr0labels_by_frame sample_dictr-rAr.__next__psD        zCharades.__next__cCs|Sr=r-r>r-r-r.__iter__szCharades.__iter__)__name__ __module__ __qualname____doc__ NUM_CLASSESrutilsdata RandomSamplerrPr rSamplerrrdictrintr/ staticmethodrr'propertyrrXrYr-r-r-r.rsD   :  Crrvideo_path_label_filerrc s ttttt|dM}tj|dd}|D];}t|dks"J|d}tj ||d}| ||d dd }g}|rLd d | d D}| |qWd n1s^wY } fdd | Dfdd | Ddd D} | fS)ap Args: video_path_label_file (List[str]): a file that contains frame paths for each video and the corresponding frame label. The file must be a space separated csv of the format: `original_vido_id video_id frame_id path labels` prefix (str): prefix path to add to all paths from video_path_label_file. r ) delimiteroriginal_vido_idpathlabels"rcSsg|]}t|qSr-)rd)r2xr-r-r.r5r6z0_read_video_paths_and_labels..,Ncr1r-r-r2key) image_pathsr-r.r5r6cr1r-r-rr)rnr-r.r5r6cSsg|] }tttj|qSr-)listset itertoolschain)r2 label_listr-r-r.r5s)rrur opencsv DictReaderr7osrmjoinappendreplacesplitkeys) rgrf csv_readerrowrErm frame_labelsry video_names video_labelsr-)rtrnr.rs*  r)r)r{r%rwr} collectionsrtypingrrrrrrrtorch.utils.dataiopath.common.file_ior pytorchvideo.data.clip_samplingr pytorchvideo.data.frame_videor r_r r`IterableDatasetrrPrdrr-r-r-r.s,      &