megadetector 5.0.11__py3-none-any.whl → 5.0.12__py3-none-any.whl

megadetector/api/batch_processing/api_support/summarize_daily_activity.py

@@ -0,0 +1,152 @@
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+"""
+This script can be run in a separate process to monitor all instances of the batch API.
+It sends a digest of submissions within the past day to a Teams channel webhook.
+
+It requires the environment variables TEAMS_WEBHOOK, COSMOS_ENDPOINT and COSMOS_READ_KEY to be set.
+"""
+
+import time
+import os
+import json
+from datetime import datetime, timezone, timedelta
+from collections import defaultdict
+
+import requests
+from azure.cosmos.cosmos_client import CosmosClient
+
+
+# Cosmos DB `batch-api-jobs` table for job status
+COSMOS_ENDPOINT = os.environ['COSMOS_ENDPOINT']
+COSMOS_READ_KEY = os.environ['COSMOS_READ_KEY']
+
+TEAMS_WEBHOOK = os.environ['TEAMS_WEBHOOK']
+
+
+def send_message():
+    cosmos_client = CosmosClient(COSMOS_ENDPOINT, credential=COSMOS_READ_KEY)
+    db_client = cosmos_client.get_database_client('camera-trap')
+    db_jobs_client = db_client.get_container_client('batch_api_jobs')
+
+    yesterday = datetime.now(timezone.utc).date() - timedelta(days=1)
+
+    query = f'''
+    SELECT *
+    FROM job
+    WHERE job.job_submission_time >= "{yesterday.isoformat()}T00:00:00Z"
+    '''
+
+    result_iterable = db_jobs_client.query_items(query=query,
+                                                 enable_cross_partition_query=True)
+
+    # aggregate the number of images, country and organization names info from each job
+    # submitted during yesterday (UTC time)
+    instance_num_images = defaultdict(lambda: defaultdict(int))
+    instance_countries = defaultdict(set)
+    instance_orgs = defaultdict(set)
+
+    total_images_received = 0
+
+    for job in result_iterable:
+        api_instance = job['api_instance']
+        status = job['status']
+        call_params = job['call_params']
+
+        if status['request_status'] == 'completed':
+            instance_num_images[api_instance]['num_images_completed'] += status.get('num_images', 0)
+        instance_num_images[api_instance]['num_images_total'] += status.get('num_images', 0)
+        total_images_received += status.get('num_images', 0)
+
+        instance_countries[api_instance].add(call_params.get('country', 'unknown'))
+        instance_orgs[api_instance].add(call_params.get('organization_name', 'unknown'))
+
+    print(f'send_message, number of images received yesterday: {total_images_received}')
+
+    if total_images_received < 1:
+        print('send_message, no images submitted yesterday, not sending a summary')
+        print('')
+        return
+
+    # create the card
+    sections = []
+
+    for instance_name, num_images in instance_num_images.items():
+        entry = {
+            'activityTitle': f'API instance: {instance_name}',
+            'facts': [
+                {
+                    'name': 'Total images',
+                    'value': '{:,}'.format(num_images['num_images_total'])
+                },
+                {
+                    'name': 'Images completed',
+                    'value': '{:,}'.format(num_images['num_images_completed'])
+                },
+                {
+                    'name': 'Countries',
+                    'value': ', '.join(sorted(list(instance_countries[instance_name])))
+                },
+                {
+                    'name': 'Organizations',
+                    'value': ', '.join(sorted(list(instance_orgs[instance_name])))
+                }
+            ]
+        }
+        sections.append(entry)
+
+    card = {
+        '@type': 'MessageCard',
+        '@context': 'http://schema.org/extensions',
+        'themeColor': 'ffcdb2',
+        'summary': 'Digest of batch API activities over the past 24 hours',
+        'title': f'Camera traps batch API activities on {yesterday.strftime("%b %d, %Y")}',
+        'sections': sections,
+        'potentialAction': [
+            {
+                '@type': 'OpenUri',
+                'name': 'View Batch account in Azure Portal',
+                'targets': [
+                    {
+                        'os': 'default',
+                        'uri': 'https://ms.portal.azure.com/#@microsoft.onmicrosoft.com/resource/subscriptions/74d91980-e5b4-4fd9-adb6-263b8f90ec5b/resourcegroups/camera_trap_api_rg/providers/Microsoft.Batch/batchAccounts/cameratrapssc/accountOverview'
+                    }
+                ]
+            }
+        ]
+    }
+
+    response = requests.post(TEAMS_WEBHOOK, data=json.dumps(card))
+    print(f'send_message, card to send:')
+    print(json.dumps(card, indent=4))
+    print(f'send_message, sent summary to webhook, response code: {response.status_code}')
+    print('')
+
+
+def main():
+    """
+    Wake up at 5 minutes past midnight UTC to send a summary of yesterday's activities if there were any.
+    Then goes in a loop to wake up and send a summary every 24 hours.
+    """
+    current = datetime.utcnow()
+    future = current.replace(day=current.day, hour=0, minute=5, second=0, microsecond=0) + timedelta(
+        days=1)  # current has been modified
+
+    current = datetime.utcnow()
+    duration = future - current
+
+    duration_hours = duration.seconds / (60 * 60)
+    print(f'Current time: {current}')
+    print(f'Will wake up at {future}, in {duration_hours} hours')
+    print('')
+
+    time.sleep(duration.seconds)
+
+    while True:
+        print(f'Woke up at {datetime.utcnow()}')
+        send_message()
+        time.sleep(24 * 60 * 60)
+
+if __name__ == '__main__':
+    main()

megadetector/api/batch_processing/integration/digiKam/setup.py

@@ -0,0 +1,6 @@
+from cx_Freeze import setup, Executable 
+  
+setup(name = "XMP Integration" , 
+      version = "3.0" , 
+      description = "XMP metadata writer" , 
+      executables = [Executable("xmp_integration.py")])

megadetector/api/batch_processing/integration/digiKam/xmp_integration.py

@@ -0,0 +1,465 @@
+#
+# xmp_integration.py
+#
+# Tools for loading MegaDetector batch API results into XMP metadata, specifically
+# for consumption in digiKam:
+#
+# https://cran.r-project.org/web/packages/camtrapR/vignettes/camtrapr2.html
+#
+    
+#%% Imports and constants
+
+import argparse
+import tkinter
+from tkinter import ttk, messagebox, filedialog
+
+import inspect
+import os
+import sys
+import json
+import pyexiv2
+import ntpath
+import threading
+import traceback
+
+from tqdm import tqdm
+from multiprocessing import Pool
+from multiprocessing.pool import ThreadPool
+from functools import partial
+
+category_mapping = {'person': 'Human', 'animal': 'Animal', 'vehicle': 'Vehicle'}
+
+
+#%% Class definitions
+
+class xmp_gui:
+    
+    root = None
+    textarea_min_threshold = None
+    textarea_status = None
+    textarea_remove_path = None
+    textarea_rename_conf = None
+    textarea_rename_cats = None
+    num_threads = 1
+    
+class xmp_integration_options:
+    
+    # Folder where images are stored
+    image_folder = None
+    
+    # .json file containing MegaDetector output
+    input_file = None
+    
+    # String to remove from all path names, typically representing a 
+    # prefix that was added during MegaDetector processing
+    remove_path = None
+    
+    # Optionally *rename* (not copy) all images that have no detections
+    # above [rename_conf] for the categories in rename_cats from x.jpg to
+    # x.check.jpg
+    rename_conf = None
+    
+    # Comma-deleted list of category names (or "all") to apply the rename_conf
+    # behavior to.
+    rename_cats = None
+    
+    # Minimum detection threshold (applies to all classes, defaults to None,
+    # i.e. 0.0
+    min_threshold = None
+    num_threads = 1
+    xmp_gui = None
+    
+    
+#%% Functions
+  
+def write_status(options,s):
+    
+    if options.xmp_gui is None:
+        return
+    options.xmp_gui.textarea_status.configure(state="normal")
+    options.xmp_gui.textarea_status.insert(tkinter.END, s + '\n')
+    options.xmp_gui.textarea_status.configure(state="disabled")
+
+    
+n_images_processed = 0
+    
+def update_xmp_metadata(categories, options, rename_cats, n_images, image):
+    """
+    Update the XMP metadata for a single image
+    """
+    
+    # Relative image path
+    filename = ''
+    
+    # Absolute image path
+    img_path = ''
+    
+    global n_images_processed
+    
+    try:
+        
+        filename = image['file']
+        if options.remove_path != None and len(options.remove_path) > 0:
+            filename = filename.replace(options.remove_path, '')
+        img_path = os.path.join(options.image_folder, filename)
+        assert os.path.isfile(img_path), 'Image {} not found'.format(img_path)
+        
+        # List of categories to write to XMP metadata
+        image_categories = []
+        
+        # Categories with above-threshold detections present for
+        # this image
+        original_image_cats = []
+        
+        # Maximum confidence for each category
+        original_image_cats_conf = {}
+        
+        for detection in image['detections']:
+            
+            cat = category_mapping[categories[detection['category']]]
+            
+            # Have we already added this to the list of categories to
+            # write out to this image?
+            if cat not in image_categories:
+                
+                # If we're supposed to compare to a threshold...
+                if len(options.min_threshold) > 0 and \
+                    options.min_threshold != None:
+                    if float(detection['conf']) > float(options.min_threshold):
+                        image_categories.append(cat)
+                        original_image_cats.append(
+                            categories[detection['category']])
+                        
+                # Else we treat *any* detection as valid...
+                else:
+                    image_categories.append(cat)
+                    original_image_cats.append(categories[detection['category']])
+
+            # Keep track of the highest-confidence detection for this class                
+            if options.min_threshold != None and \
+                len(options.min_threshold) > 0 and \
+                    detection['conf'] > \
+                        original_image_cats_conf.get(
+                            categories[detection['category']], 0):
+                            
+                original_image_cats_conf[categories[detection['category']]] = \
+                    detection['conf']
+                    
+        img = pyexiv2.Image(r'{0}'.format(img_path))
+        img.modify_xmp({'Xmp.lr.hierarchicalSubject': image_categories})
+        
+        # If we're doing the rename/.check behavior...
+        if not (options.rename_conf is None and options.rename_cats is None):
+            
+            matching_cats = set(rename_cats).intersection(set(original_image_cats))
+            is_conf_low = False
+            if options.min_threshold != None and len(options.min_threshold) > 0:
+                for matching_cat in matching_cats:
+                    if original_image_cats_conf[matching_cat] < float(options.rename_conf):
+                        is_conf_low = True
+            if options.min_threshold != None and \
+                len(options.min_threshold) > 0 and \
+                    len(image['detections']) == 0 or \
+                (len(options.rename_conf) > 0 and \
+                is_conf_low is True and \
+                    len(matching_cats) > 0):
+                        
+                parent_folder = os.path.dirname(img_path)
+                file_name = ntpath.basename(img_path)
+                manual_file_name = file_name.split('.')[0]+'_check' + '.' + file_name.split('.')[1]
+                os.rename(img_path, os.path.join(parent_folder, manual_file_name))
+                
+        if options.xmp_gui is not None:
+            
+            n_images_processed += 1
+            percentage = round((n_images_processed)/n_images*100)
+            options.xmp_gui.progress_bar['value'] = percentage
+            options.xmp_gui.root.update_idletasks()
+            options.xmp_gui.style.configure('text.Horizontal.Tprogress_bar',
+                            text='{:g} %'.format(percentage))
+                
+    except Exception as e:
+    
+        s = 'Error processing image {}: {}'.format(filename,str(e))
+        print(s)
+        traceback.print_exc()
+        write_status(options,s)
+        
+        if False:
+            
+            # Legacy code to rename files where XMP writing failed
+            parent_folder = os.path.dirname(img_path)
+            file_name = ntpath.basename(img_path)        
+            failed_file_name = file_name.split('.')[0]+'_failed' + '.' + file_name.split('.')[1]
+            os.rename(img_path, os.path.join(
+                parent_folder, failed_file_name))
+
+
+def process_input_data(options):
+    """
+    Main function to loop over images and modify XMP data
+    """
+    
+    if options.xmp_gui is not None:
+        
+        if (options.image_folder is None) or (len(options.image_folder) == 0):
+            tkinter.messagebox.showerror(title='Error', message='Image folder is not selected')
+            sys.exit()
+        if (options.input_file is None) or (len(options.input_file) == 0):
+            tkinter.messagebox.showerror(
+                title='Error', message='No MegaDetector .json file selected')
+            sys.exit()
+        options.remove_path = options.xmp_gui.textarea_remove_path.get()
+        options.rename_conf = options.xmp_gui.textarea_rename_conf.get()
+        options.rename_cats = options.xmp_gui.textarea_rename_cats.get()
+        options.num_threads = options.xmp_gui.textarea_num_threads.get()
+        options.min_threshold = options.xmp_gui.textarea_min_threshold.get()
+            
+    try:
+        
+        with open(options.input_file, 'r') as f:
+            data = f.read()
+    
+        data = json.loads(data)
+        categories = data['detection_categories']
+    
+        images = data['images']
+        n_images = len(images)
+        if not (options.rename_conf is None and options.rename_cats is None):
+            rename_cats = options.rename_cats.split(",")
+            if rename_cats[0] == 'all':
+                rename_cats = list(category_mapping.keys())
+        else:
+            rename_cats = []
+        if len(options.num_threads) > 0:
+            num_threads = int(options.num_threads)
+        else:
+            num_threads = 1
+        print(num_threads)
+        if options.xmp_gui is None:
+            func = partial(update_xmp_metadata, categories, options, rename_cats, n_images)
+            with Pool(num_threads) as p:
+                with tqdm(total=n_images) as pbar:
+                    for i, _ in enumerate(p.imap_unordered(func, images)):
+                        pbar.update()
+        else:
+            func = partial(update_xmp_metadata, categories, options, rename_cats, n_images)
+            with ThreadPool(num_threads) as p:
+                p.map(func, images)
+            s = 'Successfully processed {} images'.format(n_images)
+            print(s)
+            write_status(options,s)
+        
+    except Exception as e:
+        
+        print('Error processing input data: {}'.format(str(e)))
+        traceback.print_exc()
+        if options.xmp_gui is not None:
+            tkinter.messagebox.showerror(title='Error', 
+                                         message='Make Sure you selected the proper image folder and JSON files')
+        sys.exit()
+
+
+def start_input_processing(options):
+    
+    t = threading.Thread(target=lambda: process_input_data(options))
+    t.start()
+
+
+def browse_folder(options,folder_path_var):
+    
+    filename = tkinter.filedialog.askdirectory()
+    options.image_folder = r'{0}'.format(filename)
+    folder_path_var.set(filename)
+
+
+def browse_file(options,file_path_var):
+    
+    filename = tkinter.filedialog.askopenfilename()
+    options.input_file = r'{0}'.format(filename)
+    file_path_var.set(filename)
+
+
+def create_gui(options):
+    
+    root = tkinter.Tk()
+    root.resizable(False, False)
+    root.configure(background='white')
+    root.title('DigiKam Integration')
+    
+    group = tkinter.LabelFrame(root, padx=5, pady=5)
+    group.configure(background = 'white')
+    group.pack(padx=10, pady=10, fill='both', expand='yes')
+
+    canvas = tkinter.Canvas(group, width = 800, height = 150)    
+    canvas.configure(background = 'white')
+    canvas.pack()      
+    img1 = tkinter.PhotoImage(file='images/aiforearth.png')      
+    canvas.create_image(0,0, anchor=tkinter.NW, image=img1)
+    img2 = tkinter.PhotoImage(file='images/bg.png')      
+    canvas.create_image(0,20, anchor=tkinter.NW, image=img2)
+
+    frame = tkinter.Frame(root)
+    frame.configure(background='white')
+    frame.pack()
+    
+    l1 = tkinter.Label(frame, text='Folder containing images')
+    l1.configure(background='white')
+    l1.grid(row=0, column=0)
+    
+    folder_path_var = tkinter.StringVar()
+    
+    e1 = tkinter.Entry(frame, width=50, textvariable=folder_path_var, highlightthickness=1) 
+    e1.configure(highlightbackground='grey', highlightcolor='grey')
+    e1.grid(row=0, column=2)
+    
+    b1 = tkinter.Button(frame, text='Browse', fg='blue', command=lambda: browse_folder(options,folder_path_var))
+    b1.grid(row=0, column=5, padx=10)
+    
+    l2 = tkinter.Label(frame, text='Path to MegaDetector output .json file') 
+    l2.configure(background='white')
+    l2.grid(row=1, column=0)
+    
+    file_path_var = tkinter.StringVar()
+    
+    e2 = tkinter.Entry(frame, width=50, textvariable=file_path_var, highlightthickness=1)
+    e2.configure(highlightbackground='grey', highlightcolor='grey')    
+    e2.grid(row=1, column=2)
+    
+    b2 = tkinter.Button(frame, text='Browse', fg='blue', command=lambda: browse_file(options,file_path_var))
+    b2.grid(row=1, column=5, padx=10)
+
+    l6 = tkinter.Label(frame, text='Minimum confidence to consider a category') 
+    l6.configure(background='white')
+    l6.grid(row=2, column=0)
+    
+    textarea_min_threshold = tkinter.Entry(frame, width=50, highlightthickness=1)
+    textarea_min_threshold.configure(highlightbackground='grey', highlightcolor='grey')
+    textarea_min_threshold.grid(row=2, column=2)
+    
+    l3 = tkinter.Label(frame, text='Prefix to remove from image paths (optional)') 
+    l3.configure(background='white')
+    l3.grid(row=3, column=0)
+    
+    textarea_remove_path = tkinter.Entry(frame, width=50, highlightthickness=1)
+    textarea_remove_path.configure(highlightbackground='grey', highlightcolor='grey')
+    textarea_remove_path.grid(row=3, column=2)
+
+    l4 = tkinter.Label(frame, text='Confidence level to move images requires manual check (optional)') 
+    l4.configure(background='white')
+    l4.grid(row=4, column=0)
+
+    textarea_rename_conf = tkinter.Entry(frame, width=50, highlightthickness=1)
+    textarea_rename_conf.configure(highlightbackground='grey', highlightcolor='grey')
+    textarea_rename_conf.grid(row=4, column=2)
+
+
+    l5 = tkinter.Label(frame, text='Categories to check for the confidence (optional)') 
+    l5.configure(background='white')
+    l5.grid(row=5, column=0)
+
+    textarea_rename_cats = tkinter.Entry(frame, width=50, highlightthickness=1)
+    textarea_rename_cats.configure(highlightbackground='grey', highlightcolor='grey')
+    textarea_rename_cats.grid(row=5, column=2)
+
+    l6 = tkinter.Label(frame, text='Number of threads to run (optional)') 
+    l6.configure(background='white')
+    l6.grid(row=6, column=0)
+
+    textarea_num_threads = tkinter.Entry(frame, width=50, highlightthickness=1)
+    textarea_num_threads.configure(highlightbackground='grey', highlightcolor='grey')
+    textarea_num_threads.grid(row=6, column=2)
+    
+    sb = tkinter.Button(frame, text='Submit', fg='black',
+                command=lambda: start_input_processing(options), padx=10)
+    sb.grid(row=7, column=2, padx=10, pady=10)
+
+    style = tkinter.ttk.Style(root)
+    style.layout('text.Horizontal.Tprogress_bar',
+                [('Horizontal.progress_bar.trough',
+                {'children': [('Horizontal.progress_bar.pbar',
+                                {'side': 'left', 'sticky': 'ns'})],
+                    'sticky': 'nswe'}),
+                ('Horizontal.progress_bar.label', {'sticky': ''})])    
+    style.configure('text.Horizontal.Tprogress_bar', text='0 %')
+
+    progress_bar = tkinter.ttk.Progressbar(root, style='text.Horizontal.Tprogress_bar', length=700,
+                                maximum=100, value=0, mode='determinate')
+    progress_bar.pack(pady=10)
+
+    group2 = tkinter.LabelFrame(root, text='Status', padx=5, pady=5)
+    group2.pack(padx=10, pady=10, fill='both', expand='yes')
+
+    textarea_status = tkinter.Text(group2, height=10, width=100)
+    textarea_status.configure(state="disabled")
+    textarea_status.pack()
+    
+    options.xmp_gui = xmp_gui()
+    options.xmp_gui.root = root
+    options.xmp_gui.textarea_min_threshold = textarea_min_threshold
+    options.xmp_gui.textarea_remove_path = textarea_remove_path
+    options.xmp_gui.textarea_rename_conf = textarea_rename_conf
+    options.xmp_gui.textarea_rename_cats = textarea_rename_cats
+    options.xmp_gui.textarea_num_threads = textarea_num_threads
+    options.xmp_gui.textarea_status = textarea_status
+    options.xmp_gui.progress_bar = progress_bar
+    options.xmp_gui.style = style
+    
+    root.mainloop()
+    
+    
+#%% Interactive/test driver
+
+if False:
+    
+    #%%
+    
+    options = xmp_integration_options()
+    options.input_file = r"C:\temp\demo_images\ssmini_xmp_test_orig\ssmini.mdv4.json"
+    options.image_folder = r"C:\temp\demo_images\ssmini_xmp_test"
+    options.remove_path = 'my_images/'
+    process_input_data(options)
+    
+    
+#%% Command-line driver
+
+def args_to_object(args,obj):
+    """
+    Copy all fields from the argparse table "args" to the object "obj"
+    """
+    for n, v in inspect.getmembers(args):
+        if not n.startswith('_'):
+            setattr(obj, n, v)
+
+
+def main():
+    
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--input_file', help = 'Path to the MegaDetector .json file', default=None)
+    parser.add_argument('--image_folder', help = 'Path to the folder containing images', default=None)
+    parser.add_argument('--min_threshold', help = 'Minimum detection confidence that will be treated as a detection event', default=None)
+    parser.add_argument('--remove_path', help = 'Prefix to remove from image paths in the .json file', default=None)
+    parser.add_argument('--rename_conf', help = 'Below this confidence level, images will be renamed for manual check', default=None)
+    parser.add_argument('--rename_cat', help = 'Category (or comma-delimited categories) to apply renaming behavior to', default=None)
+    parser.add_argument('--num_threads', help = 'Number of threads to use for image processing', default=1)
+    parser.add_argument('--gui', help = 'Run in GUI mode', action='store_true')
+    
+    options = xmp_integration_options()
+    args = parser.parse_args()
+    args_to_object(args,options)
+    
+    if options.gui:
+        assert options.input_file is None, 'Command-line argument specified in GUI mode'
+        assert options.image_folder is None, 'Command-line argument specified in GUI mode'
+        assert options.min_threshold is None, 'Command-line argument specified in GUI mode'
+        assert options.remove_path is None, 'Command-line argument specified in GUI mode'
+        assert options.rename_conf is None, 'Command-line argument specified in GUI mode'
+        assert options.rename_cat is None, 'Command-line argument specified in GUI mode'
+        assert options.num_threads == 1, 'Command-line argument specified in GUI mode'
+        create_gui(options)    
+    else:
+        process_input_data(options)
+
+if __name__ == '__main__':
+    
+    main()

megadetector/api/batch_processing/integration/eMammal/test_scripts/config_template.py

@@ -0,0 +1,5 @@
+host="localhost"
+username="root"
+password=""
+port=3307
+database="wild_id"