Remove MMOCR

app_pages/ocr_comparator.py

@@ -2,12 +2,12 @@
    EasyOcr, PaddleOCR, MMOCR, Tesseract
 """
 
-import mim
-
-mim.install(['mmengine>=0.7.1,<1.1.0'])
-mim.install(['mmcv>=2.0.0rc4,<2.1.0'])
-mim.install(['mmdet>=3.0.rc5,<3.2.0'])
-mim.install(['mmocr'])
+#import mim
+#
+#mim.install(['mmengine>=0.7.1,<1.1.0'])
+#mim.install(['mmcv>=2.0.0rc4,<2.1.0'])
+#mim.install(['mmdet>=3.0.rc5,<3.2.0'])
+#mim.install(['mmocr'])
 
 import streamlit as st
 import plotly.express as px
@@ -21,7 +21,7 @@ from PIL import Image, ImageColor
 import PIL
 import easyocr
 from paddleocr import PaddleOCR
-from mmocr.utils.ocr import MMOCR
+#from mmocr.utils.ocr import MMOCR
 import pytesseract
 from pytesseract import Output
 import os
@@ -80,8 +80,10 @@ def app():
             plotly figure   : confidence color scale figure
         """
         # the readers considered
-        out_reader_type_list = ['EasyOCR', 'PPOCR', 'MMOCR', 'Tesseract']
-        out_reader_type_dict = {'EasyOCR': 0, 'PPOCR': 1, 'MMOCR': 2, 'Tesseract': 3}
+        #out_reader_type_list = ['EasyOCR', 'PPOCR', 'MMOCR', 'Tesseract']
+        #out_reader_type_dict = {'EasyOCR': 0, 'PPOCR': 1, 'MMOCR': 2, 'Tesseract': 3}
+        out_reader_type_list = ['EasyOCR', 'PPOCR', 'Tesseract']
+        out_reader_type_dict = {'EasyOCR': 0, 'PPOCR': 1, 'Tesseract': 2}
 
         # Columns for recognition details results
         out_cols_size = [2] + [2,1]*(len(out_reader_type_list)-1) # Except Tesseract
@@ -123,7 +125,7 @@ def app():
         'Tagalog': 'tl', 'Tamil': 'ta', 'Telugu': 'te', 'Turkish': 'tr', 'Ukranian': 'uk', \
         'Urdu': 'ur', 'Uyghur': 'ug', 'Uzbek': 'uz', 'Vietnamese': 'vi', 'Welsh': 'cy'}
 
-        out_dict_lang_mmocr = {'English & Chinese': 'en'}
+        #out_dict_lang_mmocr = {'English & Chinese': 'en'}
 
         out_dict_lang_tesseract = {'Afrikaans': 'afr','Albanian': 'sqi','Amharic': 'amh', \
         'Arabic': 'ara', 'Armenian': 'hye','Assamese': 'asm','Azerbaijani - Cyrilic': 'aze_cyrl', \
@@ -156,7 +158,8 @@ def app():
         'Uzbek - Cyrilic': 'uzb_cyrl','Uzbek': 'uzb','Vietnamese': 'vie','Welsh': 'cym', \
         'Western Frisian': 'fry','Yiddish': 'yid','Yoruba': 'yor'}
 
-        out_list_dict_lang = [out_dict_lang_easyocr, out_dict_lang_ppocr, out_dict_lang_mmocr, \
+        out_list_dict_lang = [out_dict_lang_easyocr, out_dict_lang_ppocr, \
+                              #out_dict_lang_mmocr, \
                             out_dict_lang_tesseract]
 
         # Initialization of detection form
@@ -221,18 +224,18 @@ def app():
         return out_ocr
 
     ###
-    @st.experimental_memo(show_spinner=False)
-    def init_mmocr(in_params):
-        """Initialization of MMOCR reader
-
-        Args:
-            in_params (dict): dict with parameters
-
-        Returns:
-            mmocr reader: the ppocr reader instance
-        """
-        out_ocr = MMOCR(recog=None, **in_params[1])
-        return out_ocr
+    #@st.experimental_memo(show_spinner=False)
+    #def init_mmocr(in_params):
+    #    """Initialization of MMOCR reader
+#
+    #    Args:
+    #        in_params (dict): dict with parameters
+#
+    #    Returns:
+    #        mmocr reader: the ppocr reader instance
+    #    """
+    #    out_ocr = MMOCR(recog=None, **in_params[1])
+    #    return out_ocr
 
     ###
     def init_readers(in_list_params):
@@ -255,10 +258,10 @@ def app():
             reader_ppocr = init_ppocr(in_list_params[1])
 
         # - MMOCR
-        with st.spinner("MMOCR reader initialization in progress ..."):
-            reader_mmocr = init_mmocr(in_list_params[2])
+        #with st.spinner("MMOCR reader initialization in progress ..."):
+        #    reader_mmocr = init_mmocr(in_list_params[2])
 
-        out_list_readers = [reader_easyocr, reader_ppocr, reader_mmocr]
+        out_list_readers = [reader_easyocr, reader_ppocr] #, reader_mmocr]
 
         return out_list_readers
 
@@ -351,44 +354,44 @@ def app():
         return out_ppocr_boxes_coordinates, out_status
 
     ###
-    @st.experimental_memo(show_spinner=False)
-    def mmocr_detect(_in_reader, in_image_path):
-        """Detection with MMOCR
-
-        Args:
-            _in_reader (EasyORC reader) : the previously initialized instance
-            in_image_path (string)      : locally saved image path
-            in_params (list)            : list with the parameters
-
-        Returns:
-            list        : list of the boxes coordinates
-            exception on error, string 'OK' otherwise
-        """
-        # MMOCR detection method
-        out_mmocr_boxes_coordinates = []
-        try:
-            det_result = _in_reader.readtext(in_image_path, details=True)
-            bboxes_list = [res['boundary_result'] for res in det_result]
-            for bboxes in bboxes_list:
-                for bbox in bboxes:
-                    if len(bbox) > 9:
-                        min_x = min(bbox[0:-1:2])
-                        min_y = min(bbox[1:-1:2])
-                        max_x = max(bbox[0:-1:2])
-                        max_y = max(bbox[1:-1:2])
-                        #box = [min_x, min_y, max_x, min_y, max_x, max_y, min_x, max_y]
-                    else:
-                        min_x = min(bbox[0:-1:2])
-                        min_y = min(bbox[1::2])
-                        max_x = max(bbox[0:-1:2])
-                        max_y = max(bbox[1::2])
-                    box4 = [ [min_x, min_y], [max_x, min_y], [max_x, max_y], [min_x, max_y] ]
-                    out_mmocr_boxes_coordinates.append(box4)
-            out_status = 'OK'
-        except Exception as e:
-            out_status = e
-
-        return out_mmocr_boxes_coordinates, out_status
+    #@st.experimental_memo(show_spinner=False)
+    #def mmocr_detect(_in_reader, in_image_path):
+    #    """Detection with MMOCR
+#
+    #    Args:
+    #        _in_reader (EasyORC reader) : the previously initialized instance
+    #        in_image_path (string)      : locally saved image path
+    #        in_params (list)            : list with the parameters
+#
+    #    Returns:
+    #        list        : list of the boxes coordinates
+    #        exception on error, string 'OK' otherwise
+    #    """
+    #    # MMOCR detection method
+    #    out_mmocr_boxes_coordinates = []
+    #    try:
+    #        det_result = _in_reader.readtext(in_image_path, details=True)
+    #        bboxes_list = [res['boundary_result'] for res in det_result]
+    #        for bboxes in bboxes_list:
+    #            for bbox in bboxes:
+    #                if len(bbox) > 9:
+    #                    min_x = min(bbox[0:-1:2])
+    #                    min_y = min(bbox[1:-1:2])
+    #                    max_x = max(bbox[0:-1:2])
+    #                    max_y = max(bbox[1:-1:2])
+    #                    #box = [min_x, min_y, max_x, min_y, max_x, max_y, min_x, max_y]
+    #                else:
+    #                    min_x = min(bbox[0:-1:2])
+    #                    min_y = min(bbox[1::2])
+    #                    max_x = max(bbox[0:-1:2])
+    #                    max_y = max(bbox[1::2])
+    #                box4 = [ [min_x, min_y], [max_x, min_y], [max_x, max_y], [min_x, max_y] ]
+    #                out_mmocr_boxes_coordinates.append(box4)
+    #        out_status = 'OK'
+    #    except Exception as e:
+    #        out_status = e
+#
+    #    return out_mmocr_boxes_coordinates, out_status
 
     ###
     def cropped_1box(in_box, in_img):
@@ -481,14 +484,14 @@ def app():
         ##
 
         ## ------- MMOCR Text detection
-        with st.spinner('MMOCR Text detection in progress ...'):
-            mmocr_boxes_coordinates, mmocr_status = mmocr_detect(_in_list_readers[2], in_image_path)
-            # Visualization
-            if mmocr_boxes_coordinates:
-                mmocr_image_detect = draw_detected(_in_list_images[0], mmocr_boxes_coordinates, \
-                                                in_color, 'None', 3)
-            else:
-                mmocr_image_detect = mmocr_status
+        #with st.spinner('MMOCR Text detection in progress ...'):
+        #    mmocr_boxes_coordinates, mmocr_status = mmocr_detect(_in_list_readers[2], in_image_path)
+        #    # Visualization
+        #    if mmocr_boxes_coordinates:
+        #        mmocr_image_detect = draw_detected(_in_list_images[0], mmocr_boxes_coordinates, \
+        #                                        in_color, 'None', 3)
+        #    else:
+        #        mmocr_image_detect = mmocr_status
         ##
 
         ## ------- Tesseract Text detection
@@ -505,9 +508,11 @@ def app():
         ##
         #
         out_list_images = _in_list_images + [easyocr_image_detect, ppocr_image_detect, \
-                                            mmocr_image_detect, tesserocr_image_detect]
+                                           # mmocr_image_detect, \
+                                                tesserocr_image_detect]
         out_list_coordinates = [easyocr_boxes_coordinates, ppocr_boxes_coordinates, \
-                                mmocr_boxes_coordinates, tesserocr_boxes_coordinates]
+                              #  mmocr_boxes_coordinates, \
+                                    tesserocr_boxes_coordinates]
         #
 
         return out_list_images, out_list_coordinates
@@ -588,8 +593,8 @@ def app():
         list_confidence_easyocr = []
         list_text_ppocr = []
         list_confidence_ppocr = []
-        list_text_mmocr = []
-        list_confidence_mmocr = []
+        #list_text_mmocr = []
+        #list_confidence_mmocr = []
 
         # Create cropped images from detection
         list_cropped_images = get_cropped(in_boxes_coordinates, in_image_cv)
@@ -607,9 +612,9 @@ def app():
         ##
 
         # Recognize with MMOCR
-        with st.spinner('MMOCR Text recognition in progress ...'):
-            list_text_mmocr, list_confidence_mmocr, status_mmocr = \
-                mmocr_recog(list_cropped_images, in_list_dict_params[2])
+        #with st.spinner('MMOCR Text recognition in progress ...'):
+        #    list_text_mmocr, list_confidence_mmocr, status_mmocr = \
+        #        mmocr_recog(list_cropped_images, in_list_dict_params[2])
         ##
 
         # Recognize with Tesseract
@@ -624,12 +629,13 @@ def app():
                                     'confidence_easyocr': list_confidence_easyocr,
                                     'text_ppocr': list_text_ppocr,
                                     'confidence_ppocr': list_confidence_ppocr,
-                                    'text_mmocr': list_text_mmocr,
-                                    'confidence_mmocr': list_confidence_mmocr
+                                    #'text_mmocr': list_text_mmocr,
+                                    #'confidence_mmocr': list_confidence_mmocr
                                     }
                                     )
 
-        out_list_reco_status = [status_easyocr, status_ppocr, status_mmocr, status_tesseract]
+        #out_list_reco_status = [status_easyocr, status_ppocr, status_mmocr, status_tesseract]
+        out_list_reco_status = [status_easyocr, status_ppocr, status_tesseract]
 
         return out_df_results, out_df_results_tesseract, out_list_reco_status
 
@@ -711,83 +717,83 @@ def app():
         return out_list_text_ppocr, out_list_confidence_ppocr, out_status
 
     ###
-    @st.experimental_memo(suppress_st_warning=True, show_spinner=False)
-    def mmocr_recog(in_list_images, in_params):
-        """Recognition with MMOCR
-
-        Args:
-            in_list_images (list) : list of cropped images
-            in_params (dict)      : parameters for recognition
-
-        Returns:
-            list : list of recognized text
-            list : list of recognition confidence
-            string/Exception : recognition status
-        """
-        ## ------- MMOCR Text recognition
-        out_list_text_mmocr = []
-        out_list_confidence_mmocr = []
-        try:
-            reader_mmocr = MMOCR(det=None, **in_params)
-            step = 2*len(in_list_images) # third recognition process
-            nb_steps = 4 * len(in_list_images)
-            progress_bar = st.progress(step/nb_steps)
-
-            for ind_img, cropped in enumerate(in_list_images):
-                result = reader_mmocr.readtext(cropped, details=True)
-                try:
-                    out_list_text_mmocr.append(result[0]['text'])
-                    out_list_confidence_mmocr.append(np.round(100* \
-                                                            (np.array(result[0]['score']).mean()), 1))
-                except:
-                    out_list_text_mmocr.append('Not recognize')
-                    out_list_confidence_mmocr.append(100.)
-                progress_bar.progress((step+ind_img+1)/nb_steps)
-            out_status = 'OK'
-        except Exception as e:
-            out_status = e
-        progress_bar.empty()
-
-        return out_list_text_mmocr, out_list_confidence_mmocr, out_status
-
-    ###
-    @st.experimental_memo(suppress_st_warning=True, show_spinner=False)
-    def tesserocr_recog(in_img, in_params, in_nb_images):
-        """Recognition with Tesseract
-
-        Args:
-            in_image_cv (matrix) : original image
-            in_params (dict)     : parameters for recognition
-            in_nb_images         : nb cropped images (used for progress bar)
-
-        Returns:
-            Pandas data frame : recognition results
-            string/Exception  : recognition status
-        """
-        ## ------- Tesseract Text recognition
-        step = 3*in_nb_images # fourth recognition process
-        nb_steps = 4 * in_nb_images
-        progress_bar = st.progress(step/nb_steps)
-
-        try:
-            out_df_result = pytesseract.image_to_data(in_img, **in_params,output_type=Output.DATAFRAME)
-
-            out_df_result['box'] = out_df_result.apply(lambda d: [[d['left'], d['top']], \
-                                                        [d['left'] + d['width'], d['top']], \
-                                                        [d['left']+d['width'], d['top']+d['height']], \
-                                                        [d['left'], d['top'] + d['height']], \
-                                                        ], axis=1)
-            out_df_result['cropped'] = out_df_result['box'].apply(lambda b: cropped_1box(b, in_img))
-            out_df_result = out_df_result[(out_df_result.word_num > 0) & (out_df_result.text != ' ')] \
-                                .reset_index(drop=True)
-            out_status = 'OK'
-        except Exception as e:
-            out_df_result = pd.DataFrame([])
-            out_status = e
-
-        progress_bar.progress(1.)
-
-        return out_df_result, out_status
+    #@st.experimental_memo(suppress_st_warning=True, show_spinner=False)
+    #def mmocr_recog(in_list_images, in_params):
+    #    """Recognition with MMOCR
+#
+    #    Args:
+    #        in_list_images (list) : list of cropped images
+    #        in_params (dict)      : parameters for recognition
+#
+    #    Returns:
+    #        list : list of recognized text
+    #        list : list of recognition confidence
+    #        string/Exception : recognition status
+    #    """
+    #    ## ------- MMOCR Text recognition
+    #    out_list_text_mmocr = []
+    #    out_list_confidence_mmocr = []
+    #    try:
+    #        reader_mmocr = MMOCR(det=None, **in_params)
+    #        step = 2*len(in_list_images) # third recognition process
+    #        nb_steps = 4 * len(in_list_images)
+    #        progress_bar = st.progress(step/nb_steps)
+#
+    #        for ind_img, cropped in enumerate(in_list_images):
+    #            result = reader_mmocr.readtext(cropped, details=True)
+    #            try:
+    #                out_list_text_mmocr.append(result[0]['text'])
+    #                out_list_confidence_mmocr.append(np.round(100* \
+    #                                                        (np.array(result[0]['score']).mean()), 1))
+    #            except:
+    #                out_list_text_mmocr.append('Not recognize')
+    #                out_list_confidence_mmocr.append(100.)
+    #            progress_bar.progress((step+ind_img+1)/nb_steps)
+    #        out_status = 'OK'
+    #    except Exception as e:
+    #        out_status = e
+    #    progress_bar.empty()
+#
+    #    return out_list_text_mmocr, out_list_confidence_mmocr, out_status
+#
+    ####
+    #@st.experimental_memo(suppress_st_warning=True, show_spinner=False)
+    #def tesserocr_recog(in_img, in_params, in_nb_images):
+    #    """Recognition with Tesseract
+#
+    #    Args:
+    #        in_image_cv (matrix) : original image
+    #        in_params (dict)     : parameters for recognition
+    #        in_nb_images         : nb cropped images (used for progress bar)
+#
+    #    Returns:
+    #        Pandas data frame : recognition results
+    #        string/Exception  : recognition status
+    #    """
+    #    ## ------- Tesseract Text recognition
+    #    step = 3*in_nb_images # fourth recognition process
+    #    nb_steps = 4 * in_nb_images
+    #    progress_bar = st.progress(step/nb_steps)
+#
+    #    try:
+    #        out_df_result = pytesseract.image_to_data(in_img, **in_params,output_type=Output.DATAFRAME)
+#
+    #        out_df_result['box'] = out_df_result.apply(lambda d: [[d['left'], d['top']], \
+    #                                                    [d['left'] + d['width'], d['top']], \
+    #                                                    [d['left']+d['width'], d['top']+d['height']], \
+    #                                                    [d['left'], d['top'] + d['height']], \
+    #                                                    ], axis=1)
+    #        out_df_result['cropped'] = out_df_result['box'].apply(lambda b: cropped_1box(b, in_img))
+    #        out_df_result = out_df_result[(out_df_result.word_num > 0) & (out_df_result.text != ' ')] \
+    #                            .reset_index(drop=True)
+    #        out_status = 'OK'
+    #    except Exception as e:
+    #        out_df_result = pd.DataFrame([])
+    #        out_status = e
+#
+    #    progress_bar.progress(1.)
+#
+    #    return out_df_result, out_status
 
     ###
     def draw_reco_images(in_image, in_boxes_coordinates, in_list_texts, in_list_confid, \
@@ -939,13 +945,13 @@ def app():
         # Clear caches
         easyocr_detect.clear()
         ppocr_detect.clear()
-        mmocr_detect.clear()
+        #mmocr_detect.clear()
         tesserocr_detect.clear()
         process_detect.clear()
         get_cropped.clear()
         easyocr_recog.clear()
         ppocr_recog.clear()
-        mmocr_recog.clear()
+        #mmocr_recog.clear()
         tesserocr_recog.clear()
 
 
@@ -953,6 +959,7 @@ def app():
     #print("PID : ", os.getpid())
 
     st.title("OCR solutions comparator")
+    #st.markdown("##### *EasyOCR, PPOCR, Tesseract*")
     st.markdown("##### *EasyOCR, PPOCR, MMOCR, Tesseract*")
     #st.markdown("#### PID : " + str(os.getpid()))
 
@@ -969,8 +976,8 @@ def app():
     easyocr_lang = list_dict_lang[0][easyocr_key_lang]
     ppocr_key_lang = lang_col[1].selectbox(reader_type_list[1]+" :", list_dict_lang[1].keys(), 22)
     ppocr_lang = list_dict_lang[1][ppocr_key_lang]
-    mmocr_key_lang = lang_col[2].selectbox(reader_type_list[2]+" :", list_dict_lang[2].keys(), 0)
-    mmocr_lang = list_dict_lang[2][mmocr_key_lang]
+    #mmocr_key_lang = lang_col[2].selectbox(reader_type_list[2]+" :", list_dict_lang[2].keys(), 0)
+    #mmocr_lang = list_dict_lang[2][mmocr_key_lang]
     tesserocr_key_lang = lang_col[3].selectbox(reader_type_list[3]+" :", list_dict_lang[3].keys(), 35)
     tesserocr_lang = list_dict_lang[3][tesserocr_key_lang]
 
@@ -1076,7 +1083,7 @@ def app():
                         help='''slow: use polygon box to calculate bbox score, fast: use rectangle box \
                         to calculate. (default = fast) \n
     Use rectlar box to calculate faster, and polygonal box more accurate for curved text area.''')
-
+            """
             with col2.expander("Choose detection hyperparameters for " + reader_type_list[2], \
                             expanded=False):
                 t2_det = st.selectbox('det', ['DB_r18','DB_r50','DBPP_r50','DRRG','FCE_IC15', \
@@ -1088,8 +1095,9 @@ def app():
                         [here](https://mmocr.readthedocs.io/en/latest/textdet_models.html)")
                 t2_merge_xdist = st.slider('merge_xdist', 1, 50, 20, step=1, \
                         help='The maximum x-axis distance to merge boxes. (defaut=20)')
-
-            with col2.expander("Choose detection hyperparameters for " + reader_type_list[3], \
+            """
+            #with col2.expander("Choose detection hyperparameters for " + reader_type_list[3], \
+            with col2.expander("Choose detection hyperparameters for " + reader_type_list[2], \
                             expanded=False):
                 t3_psm = st.selectbox('Page segmentation mode (psm)', \
                                     [' -  Default', \
@@ -1153,7 +1161,7 @@ def app():
                 'det_east_cover_thresh': t1_det_east_cover_thresh, \
                 'det_east_nms_thresh': t1_det_east_nms_thresh, \
                 'det_db_score_mode': t1_det_db_score_mode}],
-                [mmocr_lang, {'det': t2_det, 'merge_xdist': t2_merge_xdist}],
+                #[mmocr_lang, {'det': t2_det, 'merge_xdist': t2_merge_xdist}],
                 [tesserocr_lang, {'lang': tesserocr_lang, 'config': t3_config}]
                 ]
 
@@ -1264,7 +1272,7 @@ def app():
                     t1_drop_score = st.slider('drop_score', 0., 1., 0.25, step=.05, \
                         help="Filter the output by score (from the recognition model), and those \
                             below this score will not be returned. (default=0.5)")
-
+                """
                 with st.expander("Choose recognition hyperparameters for " + reader_type_list[2], \
                                 expanded=False):
                     t2_recog = st.selectbox('recog', ['ABINet','CRNN','CRNN_TPS','MASTER', \
@@ -1273,8 +1281,9 @@ def app():
                             help='Text recognition algorithm. (default = SAR)')
                     st.write("###### *More about text recognition models*  👉  \
                             [here](https://mmocr.readthedocs.io/en/latest/textrecog_models.html)")
-
-                with st.expander("Choose recognition hyperparameters for " + reader_type_list[3], \
+                """
+                #with st.expander("Choose recognition hyperparameters for " + reader_type_list[3], \
+                with st.expander("Choose recognition hyperparameters for " + reader_type_list[2], \
                                 expanded=False):
                     t3r_psm = st.selectbox('Page segmentation mode (psm)', \
                                         [' -  Default', \
@@ -1387,7 +1396,8 @@ def app():
                                         in_conf_threshold=st.session_state.conf_threshold_sld)
 
                     st.subheader("Recognition details")
-                    with st.expander("Detailed areas for EasyOCR, PPOCR, MMOCR", expanded=True):
+                    #with st.expander("Detailed areas for EasyOCR, PPOCR, MMOCR", expanded=True):
+                    with st.expander("Detailed areas for EasyOCR, PPOCR", expanded=True):
                         cols = st.columns(cols_size)
                         cols[0].markdown('#### Detected area')
                         for i in range(1, (len(reader_type_list)-1)*2, 2):