khddlZddlZddlZddlZddlmZmZddlmZddl m Z e rddl Z GddeZ Gd d eZ y) N) BaseOCRAgentBaseOCRElementType)load_dataframe)is_pytesseract_availablecDeZdZdZdZdZdZdZdZe dZ e dZ y ) TesseractFeatureTypez7 The element types for Tesseract Detection API rrrc tjdtjdtjdtjdtj di}||S)Npage_num block_numpar_numline_numword_num)r PAGEBLOCKPARALINEWORD)selfname_cvts \/var/www/teggl/fontify/venv/lib/python3.12/site-packages/layoutparser/ocr/tesseract_agent.py attr_namezTesseractFeatureType.attr_name(sN ! % %z & & % %y % %z % %z  ~cgd}|d|dzS)N)rrrrrr)rlevelss r group_levelsz!TesseractFeatureType.group_levels3sMjq!!rN) __name__ __module__ __qualname____doc__rrrrrpropertyrr rrrr r sH D E D D D ""rr cneZdZdZdgZd dZedZdZ d dZ e dZ e d Z e d Z y) TesseractAgentz A wrapper for `Tesseract `_ Text Detection APIs based on `PyTesseract `_. pytesseractc bt|tr|ndj||_||_y)afCreate a Tesseract OCR Agent. Args: languages (:obj:`list` or :obj:`str`, optional): You can specify the language code(s) of the documents to detect to improve accuracy. The supported language and their code can be found on `its github repo `_. It supports two formats: 1) you can pass in the languages code as a string of format like `"eng+fra"`, or 2) you can pack them as a list of strings `["eng", "fra"]`. Defaults to 'eng'. +N) isinstancestrjoinlangconfigs)r languageskwargss r__init__zTesseractAgent.__init__As&",Is!;I)AT  rc <|tj_|di|S)Nr)r( tesseract_cmd)clstesseract_cmd_pathr1s rwith_tesseract_executablez(TesseractAgent.with_tesseract_executableQs1C -}V}rcHi}tj|fd|ji|j|d<tj|fd|ji|j}t j tj|tjdddti|d<|S)Nr.textzutf-8 )quotingencodingsep convertersdata) r(image_to_stringr.r/ image_to_datapdread_csvioStringIOcsv QUOTE_NONEr,)r img_contentres_datas r_detectzTesseractAgent._detectWs!11  "ii +/<< F ))+VDIIVVkk KK NN}  F  rNcl|j|}|r|S|r|dS||j||S|dS)aSend the input image for OCR. Args: image (:obj:`np.ndarray` or :obj:`str`): The input image array or the name of the image file return_response (:obj:`bool`, optional): Whether directly return all output (string and boxes info) from Tesseract. Defaults to `False`. return_only_text (:obj:`bool`, optional): Whether return only the texts in the OCR results. Defaults to `False`. agg_output_level (:obj:`~TesseractFeatureType`, optional): When set, aggregate the GCV output with respect to the specified aggregation level. Defaults to `None`. r9)rK gather_data)rimagereturn_responsereturn_only_textagg_output_levelrIs rdetectzTesseractAgent.detectfsK(ll5! J v;   '##C)9: :6{rc t|ts Jd||d}||jjj |j j djdjjdddd d d d d jdddjdd g}t|S)zo Gather the OCR'ed text, bounding boxes, and confidence in a given aggeragation level. zInvalid agg_level r?c tj|dj|dj|dj|dj|dj |dj j dgS) Nlefttopwidthheightconfr9 )r=)rBSeriesminmaxmeanr,cat)gps rz,TesseractAgent.gather_data..st2996 (5 7 )8 ((*6 )6 **s*3  rT)dropx_1y_1whscorer9id)rrrr r index)columnsc4|j|jzSN)rcrexs rraz,TesseractAgent.gather_data..aeeacckrc4|j|jzSrm)rdrfrns rraz,TesseractAgent.gather_data..rpr rectangle)x_2y_2 block_type) r+r r9isnagroupbyr apply reset_indexrenameassignrbr)response agg_levelrIdfs rrMzTesseractAgent.gather_datas  +  ,  { + , v  ! WY++ , U  [d[ # [] V! V))& T3*T %C Hb!!rctt|d5}tj|}ddd|S#1swYSxYw)Nrb)openpickleload)filenamefprIs r load_responsezTesseractAgent.load_responses5 (D ! "R++b/C "  " s-7ct|d5}tj||tjdddy#1swYyxYw)Nwb)protocol)rrdumpHIGHEST_PROTOCOL)rI file_namers r save_responsezTesseractAgent.save_responses?)T " Cb KKR&*A*A B C C Cs '=A)eng)FTN)r!r"r#r$ DEPENDENCIESr2 classmethodr7rKrR staticmethodrMrrrrrr'r'9sz "?L   UYB-"-"^ CCrr')rDrFrpandasrBbaserrr file_utilsrr(r r'rrrrsB 21"-"8HC\HCr