pyppbox 3.3b1__tar.gz → 3.4b2__tar.gz

{pyppbox-3.3b1 → pyppbox-3.4b2}/GETSTARTED.md

@@ -90,11 +90,11 @@ You need to install the main package which is `pyppbox` and the data for the mod
   - Download the latest from [releases](https://github.com/rathaumons/pyppbox-data/releases)
   - Or install the ones you need directly:
     ```
-    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.0/pyppbox_data_yolocls-1.1.0-py3-none-any.whl
-    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.0/pyppbox_data_yoloult-1.1.0-py3-none-any.whl
-    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.0/pyppbox_data_deepsort-1.1.0-py3-none-any.whl
-    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.0/pyppbox_data_facenet-1.1.0-py3-none-any.whl
-    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.0/pyppbox_data_torchreid-1.1.0-py3-none-any.whl
+    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.1/pyppbox_data_yolocls-1.1.1-py3-none-any.whl
+    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.1/pyppbox_data_yoloult-1.1.1-py3-none-any.whl
+    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.1/pyppbox_data_deepsort-1.1.1-py3-none-any.whl
+    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.1/pyppbox_data_facenet-1.1.1-py3-none-any.whl
+    pip install https://github.com/rathaumons/pyppbox-data/releases/download/v1.1.1/pyppbox_data_torchreid-1.1.1-py3-none-any.whl
     ```
 
 * Install [`pyppbox-data-gta5`](https://github.com/rathaumons/PoseTReID_DATASET#-introducing-pyppbox-data-gta5)

{pyppbox-3.3b1/pyppbox.egg-info → pyppbox-3.4b2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: pyppbox
-Version: 3.3b1
+Version: 3.4b2
 Summary: Toolbox for people detecting, tracking, and re-identifying.
 Home-page: https://github.com/rathaumons/pyppbox
 Author: Ratha SIV
@@ -74,7 +74,7 @@ Requires-Dist: pyppbox-ultralytics>=8.0.218
 
 </div>
 
-***pyppbox*** is an all-in-one Python toolbox which can be used for detecting, tracking, and re-identifying people with only a few lines of codes. It is originally made for [***PoseTReID framework***](https://github.com/rathaumons/PoseTReID_DATASET) which can effectively track specific/certain people across multiple cameras or video scenes in distributed contexts of people interaction spaces such as malls or amusement parks, etc.
+***pyppbox*** is an all-in-one Python toolbox which can be used for detecting, tracking, and re-identifying people with only a few lines of code. It is originally made for [***PoseTReID framework***](https://github.com/rathaumons/PoseTReID_DATASET) which can effectively track specific/certain people across multiple cameras or video scenes in distributed contexts of people interaction spaces such as malls or amusement parks, etc.
 
 * ` pyppbox = Python + People + Toolbox `
 * Design for both short and long term people detecting, tracking, and re-identifying.

{pyppbox-3.3b1 → pyppbox-3.4b2}/README.md

@@ -12,7 +12,7 @@
 
 </div>
 
-***pyppbox*** is an all-in-one Python toolbox which can be used for detecting, tracking, and re-identifying people with only a few lines of codes. It is originally made for [***PoseTReID framework***](https://github.com/rathaumons/PoseTReID_DATASET) which can effectively track specific/certain people across multiple cameras or video scenes in distributed contexts of people interaction spaces such as malls or amusement parks, etc.
+***pyppbox*** is an all-in-one Python toolbox which can be used for detecting, tracking, and re-identifying people with only a few lines of code. It is originally made for [***PoseTReID framework***](https://github.com/rathaumons/PoseTReID_DATASET) which can effectively track specific/certain people across multiple cameras or video scenes in distributed contexts of people interaction spaces such as malls or amusement parks, etc.
 
 * ` pyppbox = Python + People + Toolbox `
 * Design for both short and long term people detecting, tracking, and re-identifying.

{pyppbox-3.3b1 → pyppbox-3.4b2}/RELEASENOTES.md

@@ -2,6 +2,26 @@
 
 ## **pyppbox V3 - Make Simpler and Faster**
 
+* `pyppbox` [v3.4b2](https://github.com/rathaumons/pyppbox/tree/v3.4b2)
+
+  - Add freedom of input video without GT (Ground-truth) in GUI demo
+  - Update GitHub workflows
+  - Update and improve documentations
+  - **Known issue/limitation**:
+    - You tell me :)
+
+* `pyppbox` [v3.4b1](https://github.com/rathaumons/pyppbox/tree/v3.4b1)
+
+  - Remove `findRepspointList()` from `pyppbox.utils.persontools`
+  - Add `findRepspointBB()` and `findRepspointUP()` to `pyppbox.utils.persontools`
+  - Add float support for the private `convertStringToNPL()` of `pyppbox.utils.gttools`
+  - Add `min_width_filter` support for `detectPeople()` of `ppbox.standalone`
+  - Add alternative repspoint support for `detectPeople()` of `ppbox.standalone` and all detectors
+  - Add [MOT Challenge ground-truth](https://motchallenge.net/instructions/) converter -> `pyppbox.utils.mot2pyppbox`
+  - Improve documentations
+  - **Known issue/limitation**:
+    - You tell me :)
+
 * `pyppbox` [v3.3b1](https://github.com/rathaumons/pyppbox/tree/v3.3b1)
 
   - Remove `ontracked` and its related methods from `pyppbox.utils.persontools.Person`

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/__init__.py

@@ -47,7 +47,7 @@ from pyppbox.gui.guitools import (
 )
 
 
-__version__ = '3.3b1'
+__version__ = '3.4b2'
 __author__ = 'Ratha SIV'
 __description__ = 'Toolbox for people detecting, tracking, and re-identifying.'
 __homepage__ = 'https://rathaumons.github.io/pyppbox'

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/config/unifiedstrings.py

@@ -46,9 +46,9 @@ class UnifiedStrings(object):
     gt : str, auto
         Unified string of words 'Ground-truth'.
     yolo_cls : str, auto
-        Unified string of wordS 'Yolo Classic'.
+        Unified string of words 'Yolo Classic'.
     yolo_ult : str, auto
-        Unified string of wordS 'Yolo Ultralytics'.
+        Unified string of words 'Yolo Ultralytics'.
     sort : str, auto
         Unified string of word 'SORT'.
     deepsort : str, auto
@@ -60,9 +60,9 @@ class UnifiedStrings(object):
     torchreid : str, auto
         Unified string of word 'Torchreid'.
     dtname_yl : str, auto
-        Unified string of wordS 'Detector YOLO'.
+        Unified string of words 'Detector YOLO'.
     dtname_gt : str, auto
-        Unified string of wordS 'Detector GT'.
+        Unified string of words 'Detector GT'.
     tkname_ct : str, auto
         Unified string of words 'Tracker Centroid'.
     tkname_st : str, auto

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/modules/detectors/yolocls/__init__.py

@@ -21,7 +21,7 @@
 
 import cv2
 
-from pyppbox.utils.persontools import Person, findRepspoint
+from pyppbox.utils.persontools import Person, findRepspoint, findRepspointBB
 from pyppbox.utils.commontools import to_xyxy
 
 
@@ -108,7 +108,7 @@ class MyYOLOCLS(object):
                                       (box_xyxy[2], box_xyxy[3]), (255, 255, 0), 2)
         return img, pboxes_xywh, pboxes_xyxy, repspoints, confs
 
-    def detectPeople(self, img, visual=True, min_width_filter=35):
+    def detectPeople(self, img, visual=True, min_width_filter=35, alt_repspoint=False, alt_repspoint_top=True):
         """Detect person(s) in a given cv :obj:`Mat` image.
 
         Parameters
@@ -116,9 +116,13 @@ class MyYOLOCLS(object):
         img : Mat
             A cv :obj:`Mat` image.
         visual : bool, default=True
-            An indication of whether to visualize the detected objects.
+            An indication of whether to visualize the detected people.
         min_width_filter : int, default=35
-            Mininum width filter of a detected object.
+            Mininum width filter of a detected person.
+        alt_repspoint : bool, default=False
+            An indication of whether to use the alternative :meth:`findRepspointBB`.
+        alt_repspoint_top : bool, default=True
+            A parameter passed to :obj:`prefer_top` of :meth:`findRepspointBB`.
 
         Returns
         -------
@@ -139,7 +143,8 @@ class MyYOLOCLS(object):
                 if class_id == 0 and box_xywh[2] >= min_width_filter:
                     box_xywh = box_xywh.astype(int)
                     box_xyxy = to_xyxy(box_xywh)
-                    repspoint = findRepspoint(box_xyxy, self.cfg.repspoint_calibration)
+                    if alt_repspoint: repspoint = findRepspointBB(box_xyxy, prefer_top=alt_repspoint_top)
+                    else: repspoint = findRepspoint(box_xyxy, self.cfg.repspoint_calibration)
                     people.append(Person(i, i, box_xywh=box_xywh, box_xyxy=box_xyxy, 
                                          repspoint=repspoint, det_conf=float(conf)))
                     i += 1

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/modules/detectors/yoloult/__init__.py

@@ -21,7 +21,7 @@
 
 import cv2
 
-from pyppbox.utils.persontools import Person, findRepspoint
+from pyppbox.utils.persontools import Person, findRepspoint, findRepspointBB
 from pyppbox.utils.commontools import to_xywh
 from pyppbox.utils.logtools import ignore_this_logger
 
@@ -204,7 +204,7 @@ class MyYOLOULT(object):
                                       (box_xyxy[2], box_xyxy[3]), (255, 255, 0), 2)
         return img, pboxes_xywh, pboxes_xyxy, repspoints, keypoints, confs
 
-    def detectPeople(self, img, visual=True, min_width_filter=35):
+    def detectPeople(self, img, visual=True, min_width_filter=35, alt_repspoint=False, alt_repspoint_top=True):
         """Detect person(s) in a given cv :obj:`Mat` image.
 
         Parameters
@@ -212,9 +212,13 @@ class MyYOLOULT(object):
         img : Mat
             A cv :obj:`Mat` image.
         visual : bool, default=True
-            An indication of whether to visualize the detected objects.
+            An indication of whether to visualize the detected people.
         min_width_filter : int, default=35
-            Mininum width filter of a detected object.
+            Mininum width filter of a detected person.
+        alt_repspoint : bool, default=False
+            An indication of whether to use the alternative :meth:`findRepspointBB`.
+        alt_repspoint_top : bool, default=True
+            A parameter passed to :obj:`prefer_top` of :meth:`findRepspointBB`.
 
         Returns
         -------
@@ -250,7 +254,8 @@ class MyYOLOULT(object):
                 box_xywh = to_xywh(box_xyxy)
                 if box_xywh[2] >= min_width_filter:
                     keypoint = kp.data[0]
-                    repspoint = findRepspoint(box_xyxy, self.cfg.repspoint_calibration)
+                    if alt_repspoint: repspoint = findRepspointBB(box_xyxy, prefer_top=alt_repspoint_top)
+                    else: repspoint = findRepspoint(box_xyxy, self.cfg.repspoint_calibration)
                     people.append(Person(i, i, box_xywh=box_xywh, box_xyxy=box_xyxy,
                                   keypoints=keypoint, repspoint=repspoint, det_conf=float(conf)))
                     i += 1
@@ -266,7 +271,8 @@ class MyYOLOULT(object):
                 box_xyxy = box_xyxy.astype(int)
                 box_xywh = to_xywh(box_xyxy)
                 if box_xywh[2] >= min_width_filter:
-                    repspoint = findRepspoint(box_xyxy, self.cfg.repspoint_calibration)
+                    if alt_repspoint: repspoint = findRepspointBB(box_xyxy, prefer_top=alt_repspoint_top)
+                    else: repspoint = findRepspoint(box_xyxy, self.cfg.repspoint_calibration)
                     people.append(Person(i, i, box_xywh=box_xywh, box_xyxy=box_xyxy, 
                                   repspoint=repspoint, det_conf=float(conf)))
                     i += 1

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/standalone/__init__.py

@@ -91,9 +91,23 @@ def setMainDetector(detector=""):
     """See :func:`pyppbox.standalone.mt.MT.setMainDetector`"""
     __stdmt__.setMainDetector(detector=detector)
 
-def detectPeople(img, img_is_mat=False, visual=False, save=False, save_file=""):
+def detectPeople(img, 
+                 img_is_mat=False, 
+                 visual=False, 
+                 save=False, 
+                 save_file="", 
+                 min_width_filter=35, 
+                 alt_repspoint=False, 
+                 alt_repspoint_top=True):
     """See :func:`pyppbox.standalone.mt.MT.detectPeople`"""
-    return __stdmt__.detectPeople(img, img_is_mat=img_is_mat, visual=visual, save=save, save_file=save_file)
+    return __stdmt__.detectPeople(img, 
+                                  img_is_mat=img_is_mat, 
+                                  visual=visual, 
+                                  save=save, 
+                                  save_file=save_file, 
+                                  min_width_filter=min_width_filter, 
+                                  alt_repspoint=alt_repspoint, 
+                                  alt_repspoint_top=alt_repspoint_top)
 
 def setMainTracker(tracker=""):
     """See :func:`pyppbox.standalone.mt.MT.setMainTracker`"""

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/standalone/mt.py

@@ -409,7 +409,15 @@ class MT(object):
         else:
             add_warning_log("---PYPPBOX : detector='" + str(detector) + "' is not recognized.")
 
-    def detectPeople(self, img, img_is_mat=False, visual=False, save=False, save_file=""): 
+    def detectPeople(self, 
+                     img, 
+                     img_is_mat=False, 
+                     visual=False, 
+                     save=False, 
+                     save_file="", 
+                     min_width_filter=35,
+                     alt_repspoint=False, 
+                     alt_repspoint_top=True): 
         """Detect people by giving an image. :func:`setConfigDir()` or :func:`setMainDetector()` must 
         be called in advance.
 
@@ -426,6 +434,12 @@ class MT(object):
             Decide whether to save the return :obj:`img` to a JPG file.
         save_file : str, default=""
             Indicate a path of where to save the processed image.
+        min_width_filter : int, default=35
+            Mininum width filter of a detected person.
+        alt_repspoint : bool, default=False
+            An indication of whether to use the alternative :meth:`findRepspointBB`.
+        alt_repspoint_top : bool, default=True
+            A parameter passed to :obj:`prefer_top` of :meth:`findRepspointBB`.
         
         Returns
         -------
@@ -440,7 +454,11 @@ class MT(object):
                 if not img_is_mat: img = getCVMat(img)
                 if (self.__dt_cfg__.dt_name.lower() == self.__unistrings__.yolo_cls or 
                     self.__dt_cfg__.dt_name.lower() == self.__unistrings__.yolo_ult):
-                    people, img = self.__dt__.detectPeople(img, visual=visual)
+                    people, img = self.__dt__.detectPeople(img, 
+                                                           visual=visual, 
+                                                           min_width_filter=min_width_filter, 
+                                                           alt_repspoint=alt_repspoint, 
+                                                           alt_repspoint_top=alt_repspoint_top)
                 elif self.__dt_cfg__.dt_name.lower() == self.__unistrings__.gt:
                     people, img = self.__dt__.getPeople(img, visual=visual)
                 if save:

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/utils/evatools.py

@@ -65,6 +65,7 @@ class MyEVA(object):
         self.current_frame = 0
         self.frame_to_check = 0
         self.id_mode = "deepid"
+        self.no_gt = False
 
     def setGTByGTMap(self, gt_map_txt, input_video, id_mode="deepid"):
         """Set a GT (Ground-truth) text file by give the GT mapping file :obj:`gt_map_txt` and 
@@ -96,8 +97,10 @@ class MyEVA(object):
         else:
             msg = ("MyEVA : setGTByGTMap() -> There is no GT file for the input '" + 
                    str(input_video) + "'")
-            add_error_log(msg)
-            raise ValueError(msg)
+            # add_error_log(msg)
+            # raise ValueError(msg)
+            add_warning_log(msg)
+            self.no_gt = True
 
     
     def setGTByKnownGTFile(self, gt_file, id_mode="deepid"):
@@ -125,8 +128,10 @@ class MyEVA(object):
         else:
             msg = ("MyEVA : setGTByKnownGTFile() -> The input gt_file='" + 
                    str(gt_file) + "' does not exist")
-            add_error_log(msg)
-            raise ValueError(msg)
+            # add_error_log(msg)
+            # raise ValueError(msg)
+            add_warning_log(msg)
+            self.no_gt = True
 
     def __checkInReID__(self):
         self.reid_count += 1
@@ -298,8 +303,11 @@ class MyEVA(object):
         score : float
             Score of the evaluation, between 0 and 1 -> (Total ID - Wrong ID - Missed Detection) / Total ID.
         """
-        self.score = float((self.gt_interpreter.total_detections - self.diff_count - 
-                            self.missed_detect) / self.gt_interpreter.total_detections)
+        if self.no_gt:
+            self.score = 0.0
+        else:
+            self.score = float((self.gt_interpreter.total_detections - self.diff_count - 
+                                self.missed_detect) / self.gt_interpreter.total_detections)
         if print_summary:
             msg = ("\n#####################################################################\n\n" +
                    "  Summary: \n\n" +

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/utils/gttools.py

@@ -326,7 +326,7 @@ def convertStringToNPL(input):
     input = input.replace("[", "")
     input = input.replace("]", "")
     input = " ".join(input.split())
-    input_list = list(map(int, input.split()))
+    input_list = list(map(float, input.split()))
     return np.array(input_list).astype(int)
 
 def formatLineInGTMap(line, splitter=':'):

pyppbox-3.4b2/pyppbox/utils/mot2pyppbox.py

@@ -0,0 +1,157 @@
+# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
+#                                                                           #
+#   pyppbox: Toolbox for people detecting, tracking, and re-identifying.    #
+#   Copyright (C) 2022 UMONS-Numediart                                      #
+#                                                                           #
+#   This program is free software: you can redistribute it and/or modify    #
+#   it under the terms of the GNU General Public License as published by    #
+#   the Free Software Foundation, either version 3 of the License, or       #
+#   (at your option) any later version.                                     #
+#                                                                           #
+#   This program is distributed in the hope that it will be useful,         #
+#   but WITHOUT ANY WARRANTY; without even the implied warranty of          #
+#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the           #
+#   GNU General Public License for more details.                            #
+#                                                                           #
+#   You should have received a copy of the GNU General Public License       #
+#   along with this program.  If not, see <https://www.gnu.org/licenses/>.  #
+#                                                                           #
+# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
+
+
+import argparse
+
+from pyppbox.utils.gttools import convertStringToNPL
+from pyppbox.utils.commontools import to_xyxy
+from pyppbox.utils.persontools import findRepspoint, findRepspointBB
+
+# step 0: filter conf
+def filter_conf(input_line, splitter=',', min_conf=0.25, max_conf=1.00):
+    """
+    :meta private:
+    """
+    res = False
+    spl_indexs = [i for i, char in enumerate(input_line) if char == splitter]
+    conf = float(input_line[spl_indexs[5] + 1:spl_indexs[6]])
+    if conf >= min_conf and conf <= max_conf:
+        res = True
+    return res
+
+# step 1: remove the last 4 columns (from coma 6th)
+def remove_from_splitter(input_line, splitter=',', from_spl=6, with_newline=False):
+    """
+    :meta private:
+    """
+    spl_indexs = [i for i, char in enumerate(input_line) if char == splitter]
+    res = input_line[:spl_indexs[from_spl - 1]]
+    if with_newline: res = res + "\n"
+    return res
+
+# step 2: replace 2nd colum with repspoint and readID
+def replace_column_with(input_line, splitter=',', column_index=2, with_str="(0, 0)	XXX", with_newline=False):
+    """
+    :meta private:
+    """
+    spl_indexs = [i for i, char in enumerate(input_line) if char == splitter]
+    res = input_line[0:spl_indexs[column_index - 2] + 1] + with_str + input_line[spl_indexs[column_index - 1]:]
+    if with_newline: res = res + "\n"
+    return res
+
+# step 3: finalize
+def finalize(input_line, splitter=',', find_repspoint=True, alt_repspoint=False, alt_repspoint_top=True, with_newline=False):
+    """
+    :meta private:
+    """
+    spl_indexs = [i for i, char in enumerate(input_line) if char == splitter]
+    input_line = input_line[0:spl_indexs[0]] + '\t' + input_line[spl_indexs[0] + 1:]
+    bbox = input_line[spl_indexs[2] + 1:]
+    bbox = bbox.replace(',', ' ')
+    bbox_xyxy_np = to_xyxy(convertStringToNPL(bbox))
+    res = input_line[0:spl_indexs[2]] + "\t[" + bbox + ']\t' + str(bbox_xyxy_np)
+    if find_repspoint:
+        repspoint = "(0, 0)"
+        if alt_repspoint: repspoint = findRepspointBB(bbox_xyxy_np, prefer_top=alt_repspoint_top)
+        else: repspoint = findRepspoint(bbox_xyxy_np, calibrate_weight=0.25)
+        res = res.replace("(0, 0)", str(repspoint))
+    res = res.replace("    ", " ")
+    res = res.replace("   ", " ")
+    res = res.replace("  ", " ")
+    res = res.replace("[ ", "[")
+    if with_newline: res = res + "\n"
+    return res
+
+# AIO
+def convert(input_file="det.txt", 
+            output_file="det_pyppbox.txt", 
+            splitter=',', 
+            min_conf=0.25, 
+            max_conf=1.00, 
+            find_repspoint=True, 
+            alt_repspoint=False, 
+            alt_repspoint_top=True):
+    """Convert ground-truth of MOT Challenge text file to pyppbox format. 
+
+    Parameters
+    ----------
+    input_file : str, default="det.txt"
+        Input MOT Challenge text file.
+    output_file : str, default="det_pyppbox.txt"
+        Output pyppbox text file.
+    splitter : str, default=','
+        Column or data splitter.
+    min_conf : float, default=0.25
+        Filter for minimum conf.
+    max_conf : float, default=1.00
+        Filter for maximum conf.
+    find_repspoint : bool, default=True
+        Whether to calculate repspoint.
+    alt_repspoint : bool, default=False
+        Whether to use alternative repspoint.
+    alt_repspoint_top : bool, default=True
+        Whether y is the top for alternative repspoint.
+    """
+    
+    with open(output_file, 'w') as pyppboxtxt:
+        with open(input_file, 'r') as mottxt:
+            lines = mottxt.readlines()
+            for line in lines:
+                if filter_conf(line, splitter=splitter, min_conf=min_conf, max_conf=max_conf):
+                    line = line.replace("\r", "")
+                    line = line.replace("\n", "")
+                    line = remove_from_splitter(line, splitter=splitter)
+                    line = replace_column_with(line, splitter=splitter)
+                    line = finalize(line, 
+                                    splitter=splitter, 
+                                    find_repspoint=find_repspoint, 
+                                    alt_repspoint=alt_repspoint, 
+                                    alt_repspoint_top=alt_repspoint_top, 
+                                    with_newline=True)
+                    pyppboxtxt.write(line)
+
+
+###################################################################################################
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser(description="Convert ground-truth of MOT Challenge text file to pyppbox format.")
+
+    parser.add_argument("--input-file", type=str, default="det.txt", help="Input MOT Challenge text file")
+    parser.add_argument("--output-file", type=str, default="det_pyppbox.txt", help="Output pyppbox text file")
+    parser.add_argument("--splitter", type=str, default=',', help="Column or data splitter")
+    parser.add_argument("--min-conf", type=float, default='0.25', help="Filter for minimum conf")
+    parser.add_argument("--max-conf", type=float, default='1.00', help="Filter for maximum conf")
+    parser.add_argument("--find-repspoint", type=bool, default=True, help="Whether to calculate repspoint")
+    parser.add_argument("--alt-repspoint", type=bool, default=False, help="Whether to use alternative repspoint")
+    parser.add_argument("--alt-repspoint-top", type=bool, default=True, help="Whether y is the top for alternative repspoint")
+
+    args = parser.parse_args()
+    convert(input_file=args.input_file, 
+            output_file=args.output_file, 
+            splitter=args.splitter, 
+            min_conf=args.min_conf, 
+            max_conf=args.max_conf, 
+            find_repspoint=args.find_repspoint, 
+            alt_repspoint=args.alt_repspoint, 
+            alt_repspoint_top=args.alt_repspoint_top)
+

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox/utils/persontools.py

@@ -195,26 +195,67 @@ def findRepspoint(box_xyxy, calibrate_weight):
     y = int(y_start + calibrate_weight*y_dist)
     return (x, y)
 
-def findRepspointList(box_xyxy, calibrate_weight):
+
+def findRepspointBB(box_xyxy, prefer_top=True):
     """Find respesented point :code:`(x, y)` of a :class:`Person` object by its bounding 
-    :code:`box_xyxy` of :code:`[x1, y1, x2, y2]`. The :obj:`calibrate_weight` indicates, 
-    in between :code:`min(y1, y2)` and :code:`max(y1, y2)`, where the :code:`y` is.
+    :code:`box_xyxy` of :code:`[x1, y1, x2, y2]`. :code:`x` is the middle of :code:`x1` 
+    and :code:`x2` while :code:`y` is the min or max of :code:`(y1, y2)`.
 
     Parameters
     ----------
-    box_xyxy : ndarray, optional
+    box_xyxy : ndarray
         Bounding box :code:`[x1, y1, x2, y2]`, :code:`shape=(4,)`, :code:`dtype=int`, 
         :code:`ndim=1`.
     calibrate_weight : float
         Calibration weight.
+    prefer_top : bool, default=True
+        Decide whether :code:`y` is at the top or bottom of the bounding box.
 
     Returns
     -------
-    list[int, int]
-        Respesented 2D point :code:`[x, y]`.
+    tuple(int, int)
+        Respesented 2D point :code:`(x, y)`.
     """
     x = int((box_xyxy[0] + box_xyxy[2]) / 2)
-    y_start = min(box_xyxy[1], box_xyxy[3])
-    y_dist = abs(box_xyxy[1] - box_xyxy[3])
-    y = int(y_start + calibrate_weight*y_dist)
-    return [x, y]
+    y = 0
+    if prefer_top:
+        y = min(box_xyxy[1], box_xyxy[3])
+    else:
+        y = max(box_xyxy[1], box_xyxy[3])
+    return (x, y)
+
+
+def findRepspointUP(keypoint, box_xyxy, calibrate_weight, prefer_box=True):
+    """Find respesented point :code:`(x, y)` of a :class:`Person` object by its
+    YOLOv8 pose :code:`keypoint` (17 keypoints) or by the bounding :code:`box_xyxy` 
+    of :code:`[x1, y1, x2, y2]`. The :obj:`calibrate_weight` indicates, in between 
+    :code:`min(y1, y2)` and :code:`max(y1, y2)`, where the :code:`y` is.
+
+    Parameters
+    ----------
+    keypoint : ndarray
+        17 keypoints generated by YOLOv8 (Ultralytics).
+    box_xyxy : ndarray
+        Bounding box :code:`[x1, y1, x2, y2]`, :code:`shape=(4,)`, :code:`dtype=int`, 
+        :code:`ndim=1`.
+    calibrate_weight : float
+        Calibration weight.
+    prefer_box : bool, default=True
+        Generate respesented point whether by :code:`box_xyxy` or :code:`keypoint`.
+
+    Returns
+    -------
+    tuple(int, int)
+        Respesented 2D point :code:`(x, y)`.
+    """
+    x = 0
+    y = 0
+    if len(keypoint) == 17 and not prefer_box:
+        lshd = keypoint[5]
+        rshd = keypoint[6]
+        x = int((lshd[0] + rshd[0]) / 2) 
+        y = int((lshd[1] + rshd[1]) / 2) 
+    else:
+        (x, y) = findRepspoint(box_xyxy, calibrate_weight)
+    return (x, y)
+

{pyppbox-3.3b1 → pyppbox-3.4b2/pyppbox.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: pyppbox
-Version: 3.3b1
+Version: 3.4b2
 Summary: Toolbox for people detecting, tracking, and re-identifying.
 Home-page: https://github.com/rathaumons/pyppbox
 Author: Ratha SIV
@@ -74,7 +74,7 @@ Requires-Dist: pyppbox-ultralytics>=8.0.218
 
 </div>
 
-***pyppbox*** is an all-in-one Python toolbox which can be used for detecting, tracking, and re-identifying people with only a few lines of codes. It is originally made for [***PoseTReID framework***](https://github.com/rathaumons/PoseTReID_DATASET) which can effectively track specific/certain people across multiple cameras or video scenes in distributed contexts of people interaction spaces such as malls or amusement parks, etc.
+***pyppbox*** is an all-in-one Python toolbox which can be used for detecting, tracking, and re-identifying people with only a few lines of code. It is originally made for [***PoseTReID framework***](https://github.com/rathaumons/PoseTReID_DATASET) which can effectively track specific/certain people across multiple cameras or video scenes in distributed contexts of people interaction spaces such as malls or amusement parks, etc.
 
 * ` pyppbox = Python + People + Toolbox `
 * Design for both short and long term people detecting, tracking, and re-identifying.

{pyppbox-3.3b1 → pyppbox-3.4b2}/pyppbox.egg-info/SOURCES.txt

@@ -81,6 +81,7 @@ pyppbox/utils/commontools.py
 pyppbox/utils/evatools.py
 pyppbox/utils/gttools.py
 pyppbox/utils/logtools.py
+pyppbox/utils/mot2pyppbox.py
 pyppbox/utils/persontools.py
 pyppbox/utils/restools.py
 pyppbox/utils/visualizetools.py