Simple UI to see the difference and the two images

changedetectionio/__init__.py

@@ -695,6 +695,10 @@ def changedetection_app(config=None, datastore_o=None):
     @app.route("/diff/<string:uuid>", methods=['GET'])
     @login_required
     def diff_history_page(uuid):
+        from changedetectionio import content_fetcher
+
+        newest_version_file_contents = ""
+        previous_version_file_contents = ""
 
         # More for testing, possible to return the first/only
         if uuid == 'first':
@@ -720,21 +724,28 @@ def changedetection_app(config=None, datastore_o=None):
 
         # Save the current newest history as the most recently viewed
         datastore.set_last_viewed(uuid, dates[0])
-        newest_file = watch['history'][dates[0]]
-        with open(newest_file, 'r') as f:
-            newest_version_file_contents = f.read()
 
         previous_version = request.args.get('previous_version')
-        try:
-            previous_file = watch['history'][previous_version]
-        except KeyError:
-            # Not present, use a default value, the second one in the sorted list.
-            previous_file = watch['history'][dates[1]]
+        if ('content-type' in watch and content_fetcher.supported_binary_type(watch['content-type'])):
+            template = "diff-image.html"
+        else:
+            newest_file = watch['history'][dates[0]]
+            with open(newest_file, 'r') as f:
+                newest_version_file_contents = f.read()
 
-        with open(previous_file, 'r') as f:
-            previous_version_file_contents = f.read()
+            try:
+                previous_file = watch['history'][previous_version]
+            except KeyError:
+                # Not present, use a default value, the second one in the sorted list.
+                previous_file = watch['history'][dates[1]]
 
-        output = render_template("diff.html", watch_a=watch,
+            with open(previous_file, 'r') as f:
+                previous_version_file_contents = f.read()
+
+            template = "diff.html"
+
+        output = render_template(template,
+                                 watch_a=watch,
                                  newest=newest_version_file_contents,
                                  previous=previous_version_file_contents,
                                  extra_stylesheets=extra_stylesheets,
@@ -751,6 +762,7 @@ def changedetection_app(config=None, datastore_o=None):
     @app.route("/preview/<string:uuid>", methods=['GET'])
     @login_required
     def preview_page(uuid):
+        from changedetectionio import content_fetcher
 
         # More for testing, possible to return the first/only
         if uuid == 'first':
@@ -765,14 +777,25 @@ def changedetection_app(config=None, datastore_o=None):
             return redirect(url_for('index'))
 
         newest = list(watch['history'].keys())[-1]
-        with open(watch['history'][newest], 'r') as f:
-            content = f.readlines()
+        fname = watch['history'][newest]
+
+        if ('content-type' in watch and content_fetcher.supported_binary_type(watch['content-type'])):
+            template = "preview-image.html"
+            content = fname
+        else:
+            template = "preview.html"
+            try:
+                with open(fname, 'r') as f:
+                    content = f.read()
+            except:
+                content = "Cant read {}".format(fname)
 
         output = render_template("preview.html",
                                  content=content,
                                  extra_stylesheets=extra_stylesheets,
                                  current_diff_url=watch['url'],
-                                 uuid=uuid)
+                                 uuid=uuid,
+                                 watch=watch)
         return output
 
     @app.route("/settings/notification-logs", methods=['GET'])
@@ -783,6 +806,50 @@ def changedetection_app(config=None, datastore_o=None):
                                  logs=notification_debug_log if len(notification_debug_log) else ["No errors or warnings detected"])
 
         return output
+
+
+    # render an image which contains the diff of two images
+    # We always compare the newest against whatever compare_date we are given
+    @app.route("/diff/show-image/<string:uuid>/<string:datestr>")
+    def show_single_image(uuid, datestr):
+
+        from flask import make_response
+        watch = datastore.data['watching'][uuid]
+
+        if datestr == 'None' or datestr is None:
+            datestr = list(watch['history'].keys())[0]
+
+        fname = watch['history'][datestr]
+        with open(fname, 'rb') as f:
+            resp = make_response(f.read())
+            
+        # @todo assumption here about the type, re-encode? detect?
+        resp.headers['Content-Type'] = 'image/jpeg'
+        return resp
+
+    # render an image which contains the diff of two images
+    # We always compare the newest against whatever compare_date we are given
+    @app.route("/diff/image/<string:uuid>/<string:compare_date>")
+    def render_diff_image(uuid, compare_date):
+        from changedetectionio import image_diff
+
+        from flask import make_response
+        watch = datastore.data['watching'][uuid]
+        newest = list(watch['history'].keys())[-1]
+
+        # @todo this is weird
+        if compare_date == 'None' or compare_date is None:
+            compare_date = list(watch['history'].keys())[0]
+
+        new_img = watch['history'][newest]
+        prev_img = watch['history'][compare_date]
+        img = image_diff.render_diff(new_img, prev_img)
+
+        resp = make_response(img)
+        resp.headers['Content-Type'] = 'image/jpeg'
+        return resp
+
+
     @app.route("/api/<string:uuid>/snapshot/current", methods=['GET'])
     @login_required
     def api_snapshot(uuid):

changedetectionio/content_fetcher.py

@@ -5,8 +5,9 @@ from selenium import webdriver
 from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
 from selenium.webdriver.common.proxy import Proxy as SeleniumProxy
 from selenium.common.exceptions import WebDriverException
-import urllib3.exceptions
 
+# image/jpeg etc
+supported_binary_types = ['image']
 
 class EmptyReply(Exception):
     def __init__(self, status_code, url):
@@ -51,6 +52,15 @@ class Fetcher():
 #    def return_diff(self, stream_a, stream_b):
 #        return
 
+# Assume we dont support it as binary if its not in our list
+def supported_binary_type(content_type):
+    # Not a binary thing we support? then use text (also used for JSON/XML etc)
+    # @todo - future - use regex for matching
+    if content_type and content_type.lower().strip().split('/')[0] not in (string.lower() for string in supported_binary_types):
+        return False
+
+    return True
+
 def available_fetchers():
         import inspect
         from changedetectionio import content_fetcher
@@ -156,15 +166,18 @@ class html_requests(Fetcher):
                          verify=False)
 
         # https://stackoverflow.com/questions/44203397/python-requests-get-returns-improperly-decoded-text-instead-of-utf-8
-        # Return bytes here
-        html = r.text
+
+        if not supported_binary_type(r.headers.get('Content-Type', '')):
+            content = r.text
+        else:
+            content = r.content
 
         # @todo test this
         # @todo maybe you really want to test zero-byte return pages?
-        if not r or not html or not len(html):
+        if not r or not content or not len(content):
             raise EmptyReply(url=url, status_code=r.status_code)
 
         self.status_code = r.status_code
-        self.content = html
+        self.content = content
         self.headers = r.headers
 

changedetectionio/fetch_site_status.py

@@ -55,10 +55,13 @@ class perform_site_check():
 
         changed_detected = False
         stripped_text_from_html = ""
+        fetched_md5 = ""
+
+        original_content_before_filters = False
 
         watch = self.datastore.data['watching'][uuid]
-        # Unset any existing notification error
 
+        # Unset any existing notification error
         update_obj = {'last_notification_error': False, 'last_error': False}
 
         extra_headers = self.datastore.get_val(uuid, 'headers')
@@ -92,6 +95,7 @@ class perform_site_check():
 
             fetcher = klass()
             fetcher.run(url, timeout, request_headers, request_body, request_method)
+
             # Fetching complete, now filters
             # @todo move to class / maybe inside of fetcher abstract base?
 
@@ -101,26 +105,39 @@ class perform_site_check():
             #  - Do we convert to JSON?
             # https://stackoverflow.com/questions/41817578/basic-method-chaining ?
             # return content().textfilter().jsonextract().checksumcompare() ?
-
-            is_json = fetcher.headers.get('Content-Type', '') == 'application/json'
-            is_html = not is_json
+            update_obj['content-type'] = fetcher.headers.get('Content-Type', '').lower().strip()
+            
+            # Could be 'application/json; charset=utf-8' etc
+            is_json = 'application/json' in update_obj['content-type']
+            is_text_or_html = 'text/' in update_obj['content-type'] # text/plain , text/html etc
+            is_binary = not is_text_or_html and content_fetcher.supported_binary_type(update_obj['content-type'])
             css_filter_rule = watch['css_filter']
-
             has_filter_rule = css_filter_rule and len(css_filter_rule.strip())
+
+            # Auto-detect application/json, make it reformat the JSON to something nice
             if is_json and not has_filter_rule:
                 css_filter_rule = "json:$"
                 has_filter_rule = True
 
-            if has_filter_rule:
-                if 'json:' in css_filter_rule:
-                    stripped_text_from_html = html_tools.extract_json_as_string(content=fetcher.content, jsonpath_filter=css_filter_rule)
-                    is_html = False
+            ##### CONVERT THE INPUT TO TEXT, EXTRACT THE PARTS THAT NEED TO BE FILTERED
 
-            if is_html:
+            # Dont depend on the content-type header here, maybe it's not present
+            if 'json:' in css_filter_rule:
+                is_json = True
+                rule = css_filter_rule.replace('json:', '')
+                stripped_text_from_html = html_tools.extract_json_as_string(content=fetcher.content,
+                                                                            jsonpath_filter=rule).encode('utf-8')
+                is_text_or_html = False
+                original_content_before_filters = stripped_text_from_html
+
+            if is_text_or_html:
                 # CSS Filter, extract the HTML that matches and feed that into the existing inscriptis::get_text
                 html_content = fetcher.content
-                if not fetcher.headers.get('Content-Type', '') == 'text/plain':
+                if 'text/plain' in update_obj['content-type']:
+                    stripped_text_from_html = html_content
 
+                # Assume it's HTML if it's not text/plain
+                if not 'text/plain' in update_obj['content-type']:
                     if has_filter_rule:
                         # For HTML/XML we offer xpath as an option, just start a regular xPath "/.."
                         if css_filter_rule[0] == '/':
@@ -128,35 +145,52 @@ class perform_site_check():
                         else:
                             # CSS Filter, extract the HTML that matches and feed that into the existing inscriptis::get_text
                             html_content = html_tools.css_filter(css_filter=css_filter_rule, html_content=fetcher.content)
-
                     # get_text() via inscriptis
                     stripped_text_from_html = get_text(html_content)
-                else:
-                    # Don't run get_text or xpath/css filters on plaintext
-                    stripped_text_from_html = html_content
 
-            # Re #340 - return the content before the 'ignore text' was applied
-            text_content_before_ignored_filter = stripped_text_from_html.encode('utf-8')
+                    # Extract title as title
+                    if self.datastore.data['settings']['application']['extract_title_as_title'] or watch['extract_title_as_title']:
+                        if not watch['title'] or not len(watch['title']):
+                            update_obj['title'] = html_tools.extract_element(find='title', html_content=fetcher.content)
+
+                # Re #340 - return the content before the 'ignore text' was applied
+                original_content_before_filters = stripped_text_from_html.encode('utf-8')
+
 
             # We rely on the actual text in the html output.. many sites have random script vars etc,
             # in the future we'll implement other mechanisms.
 
             update_obj["last_check_status"] = fetcher.get_last_status_code()
 
-            # If there's text to skip
-            # @todo we could abstract out the get_text() to handle this cleaner
-            text_to_ignore = watch.get('ignore_text', []) + self.datastore.data['settings']['application'].get('global_ignore_text', [])
-            if len(text_to_ignore):
-                stripped_text_from_html = self.strip_ignore_text(stripped_text_from_html, text_to_ignore)
-            else:
-                stripped_text_from_html = stripped_text_from_html.encode('utf8')
+            ######## AFTER FILTERING, STRIP OUT IGNORE TEXT
+            if is_text_or_html:
+                text_to_ignore = watch.get('ignore_text', []) + self.datastore.data['settings']['application'].get('global_ignore_text', [])
+                if len(text_to_ignore):
+                    stripped_text_from_html = self.strip_ignore_text(stripped_text_from_html, text_to_ignore)
+                else:
+                    stripped_text_from_html = stripped_text_from_html.encode('utf8')
 
+
+            ######## CALCULATE CHECKSUM FOR DIFF DETECTION
             # Re #133 - if we should strip whitespaces from triggering the change detected comparison
-            if self.datastore.data['settings']['application'].get('ignore_whitespace', False):
-                fetched_md5 = hashlib.md5(stripped_text_from_html.translate(None, b'\r\n\t ')).hexdigest()
-            else:
+            if is_text_or_html:
+                if self.datastore.data['settings']['application'].get('ignore_whitespace', False):
+                    fetched_md5 = hashlib.md5(stripped_text_from_html.translate(None, b'\r\n\t ')).hexdigest()
+                else:
+                    fetched_md5 = hashlib.md5(stripped_text_from_html).hexdigest()
+
+            if is_json:
                 fetched_md5 = hashlib.md5(stripped_text_from_html).hexdigest()
 
+            # Goal here in the future is to be able to abstract out different content type checks into their own class
+
+            if is_binary:
+                # @todo - use some actual image hash here where possible, audio hash, etc etc
+                m = hashlib.sha256()
+                m.update(fetcher.content)
+                fetched_md5 = m.hexdigest()
+                original_content_before_filters = fetcher.content
+
             # On the first run of a site, watch['previous_md5'] will be an empty string, set it the current one.
             if not len(watch['previous_md5']):
                 watch['previous_md5'] = fetched_md5
@@ -164,36 +198,30 @@ class perform_site_check():
 
             blocked_by_not_found_trigger_text = False
 
-            if len(watch['trigger_text']):
-                blocked_by_not_found_trigger_text = True
-                for line in watch['trigger_text']:
-                    # Because JSON wont serialize a re.compile object
-                    if line[0] == '/' and line[-1] == '/':
-                        regex = re.compile(line.strip('/'), re.IGNORECASE)
-                        # Found it? so we don't wait for it anymore
-                        r = re.search(regex, str(stripped_text_from_html))
-                        if r:
+            # Trigger text can apply to JSON parsed documents too
+            if is_text_or_html or is_json:
+                if len(watch['trigger_text']):
+                    blocked_by_not_found_trigger_text = True
+                    for line in watch['trigger_text']:
+                        # Because JSON wont serialize a re.compile object
+                        if line[0] == '/' and line[-1] == '/':
+                            regex = re.compile(line.strip('/'), re.IGNORECASE)
+                            # Found it? so we don't wait for it anymore
+                            r = re.search(regex, str(stripped_text_from_html))
+                            if r:
+                                blocked_by_not_found_trigger_text = False
+                                break
+
+                        elif line.lower() in str(stripped_text_from_html).lower():
+                            # We found it don't wait for it.
                             blocked_by_not_found_trigger_text = False
                             break
 
-                    elif line.lower() in str(stripped_text_from_html).lower():
-                        # We found it don't wait for it.
-                        blocked_by_not_found_trigger_text = False
-                        break
-
-
-
             if not blocked_by_not_found_trigger_text and watch['previous_md5'] != fetched_md5:
                 changed_detected = True
                 update_obj["previous_md5"] = fetched_md5
                 update_obj["last_changed"] = timestamp
 
 
-            # Extract title as title
-            if is_html:
-                if self.datastore.data['settings']['application']['extract_title_as_title'] or watch['extract_title_as_title']:
-                    if not watch['title'] or not len(watch['title']):
-                        update_obj['title'] = html_tools.extract_element(find='title', html_content=fetcher.content)
-
-
-        return changed_detected, update_obj, text_content_before_ignored_filter
+        # original_content_before_filters is returned for saving the data to disk
+        return changed_detected, update_obj, original_content_before_filters

changedetectionio/image_diff.py

@@ -0,0 +1,41 @@
+# import the necessary packages
+from skimage.metrics import structural_similarity as compare_ssim
+import argparse
+import imutils
+import cv2
+
+# From https://www.pyimagesearch.com/2017/06/19/image-difference-with-opencv-and-python/
+def render_diff(fpath_imageA, fpath_imageB):
+
+	imageA = cv2.imread(fpath_imageA)
+	imageB = cv2.imread(fpath_imageB)
+
+	# convert the images to grayscale
+	grayA = cv2.cvtColor(imageA, cv2.COLOR_BGR2GRAY)
+	grayB = cv2.cvtColor(imageB, cv2.COLOR_BGR2GRAY)
+
+	# compute the Structural Similarity Index (SSIM) between the two
+	# images, ensuring that the difference image is returned
+	(score, diff) = compare_ssim(grayA, grayB, full=True)
+	diff = (diff * 255).astype("uint8")
+	print("SSIM: {}".format(score))
+
+	# threshold the difference image, followed by finding contours to
+	# obtain the regions of the two input images that differ
+	thresh = cv2.threshold(diff, 0, 255,
+		cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)[1]
+	cnts = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL,
+		cv2.CHAIN_APPROX_SIMPLE)
+	cnts = imutils.grab_contours(cnts)
+
+	# loop over the contours
+	for c in cnts:
+		# compute the bounding box of the contour and then draw the
+		# bounding box on both input images to represent where the two
+		# images differ
+		(x, y, w, h) = cv2.boundingRect(c)
+		cv2.rectangle(imageA, (x, y), (x + w, y + h), (0, 0, 255), 2)
+		cv2.rectangle(imageB, (x, y), (x + w, y + h), (0, 0, 255), 2)
+
+	#return cv2.imencode('.jpg', imageB)[1].tobytes()
+	return cv2.imencode('.jpg', imageA)[1].tobytes()

changedetectionio/store.py

@@ -372,7 +372,9 @@ class ChangeDetectionStore:
         if not os.path.isdir(output_path):
             mkdir(output_path)
 
-        fname = "{}/{}.stripped.txt".format(output_path, uuid.uuid4())
+        suffix = "stripped.txt"
+
+        fname = "{}/{}.{}".format(output_path, uuid.uuid4(), suffix)
         with open(fname, 'wb') as f:
             f.write(contents)
             f.close()

changedetectionio/templates/diff-image.html

@@ -0,0 +1,59 @@
+{% extends 'base.html' %}
+
+{% block content %}
+
+<div id="settings">
+    <h1>Differences</h1>
+    <form class="pure-form " action="" method="GET">
+        <fieldset>
+            {% if versions|length >= 1 %}
+            <label for="diff-version">Compare newest (<span id="current-v-date"></span>) with</label>
+            <select id="diff-version" name="previous_version">
+                {% for version in versions %}
+                <option value="{{version}}" {% if version== current_previous_version %} selected="" {% endif %}>
+                    {{version}}
+                </option>
+                {% endfor %}
+            </select>
+            <button type="submit" class="pure-button pure-button-primary">Go</button>
+            {% endif %}
+        </fieldset>
+    </form>
+
+</div>
+
+<div id="diff-ui">
+  <img style="max-width: 100%" src="{{ url_for('render_diff_image', uuid=uuid, compare_date=current_previous_version) }}" />
+
+    <div>
+        <span style="width: 50%">
+            <img style="max-width: 100%" src="{{ url_for('show_single_image', uuid=uuid, datestr=newest_version_timestamp) }}" />
+        </span>
+        <span style="width: 50%">
+            <img style="max-width: 100%" src="{{ url_for('show_single_image', uuid=uuid, datestr=current_previous_version) }}" />
+        </span>
+    </div>
+</div>
+
+
+<script type="text/javascript" src="{{url_for('static_content', group='js', filename='diff.js')}}"></script>
+
+<script defer="">
+window.onload = function() {
+    /* Set current version date as local time in the browser also */
+    var current_v = document.getElementById("current-v-date");
+    var dateObject = new Date({{ newest_version_timestamp }}*1000);
+    current_v.innerHTML=dateObject.toLocaleString();
+
+    /* Convert what is options from UTC time.time() to local browser time */
+    var diffList=document.getElementById("diff-version");
+    if (typeof(diffList) != 'undefined' && diffList != null) {
+        for (var option of diffList.options) {
+          var dateObject = new Date(option.value*1000);
+          option.label=dateObject.toLocaleString();
+        }
+    }
+}
+</script>
+
+{% endblock %}

changedetectionio/templates/preview-image.html

@@ -0,0 +1,13 @@
+{% extends 'base.html' %}
+
+{% block content %}
+
+<div id="settings">
+    <h1>Current</h1>
+</div>
+
+<div id="diff-ui">
+    image goes here
+</div>
+
+{% endblock %}

changedetectionio/templates/preview.html

@@ -11,7 +11,7 @@
         <tbody>
         <tr>
             <td id="diff-col">
-                <span id="result">{% for row in content %}{{row}}{% endfor %}</span>
+                <span id="result">{{content}}</span>
             </td>
         </tr>
         </tbody>

changedetectionio/tests/test_backend.py

@@ -100,6 +100,14 @@ def test_check_basic_change_detection_functionality(client, live_server):
     # It should have picked up the <title>
     assert b'head title' in res.data
 
+
+    # be sure the HTML converter worked
+    res = client.get(url_for("preview_page", uuid="first"))
+    assert b'<html>' not in res.data
+
+    res = client.get(url_for("preview_page", uuid="first"))
+    assert b'Some initial text' in res.data
+
     #
     # Cleanup everything
     res = client.get(url_for("api_delete", uuid="all"), follow_redirects=True)

changedetectionio/tests/test_binary_fetch.py

@@ -0,0 +1,56 @@
+#!/usr/bin/python3
+
+import time
+import secrets
+from flask import url_for
+from . util import live_server_setup
+
+
+def test_binary_file_change(client, live_server):
+    with open("test-datastore/test.bin", "wb") as f:
+        f.write(secrets.token_bytes())
+
+    live_server_setup(live_server)
+
+    sleep_time_for_fetch_thread = 3
+
+    # Give the endpoint time to spin up
+    time.sleep(1)
+
+    # Add our URL to the import page
+    test_url = url_for('test_binaryfile_endpoint', _external=True)
+    res = client.post(
+        url_for("import_page"),
+        data={"urls": test_url},
+        follow_redirects=True
+    )
+    assert b"1 Imported" in res.data
+
+    # Trigger a check
+    client.get(url_for("api_watch_checknow"), follow_redirects=True)
+
+    # Give the thread time to pick it up
+    time.sleep(sleep_time_for_fetch_thread)
+
+    # Trigger a check
+    client.get(url_for("api_watch_checknow"), follow_redirects=True)
+
+    # It should report nothing found (no new 'unviewed' class)
+    res = client.get(url_for("index"))
+    assert b'unviewed' not in res.data
+    assert b'/test-binary-endpoint' in res.data
+
+    #  Make a change
+    with open("test-datastore/test.bin", "wb") as f:
+        f.write(secrets.token_bytes())
+
+
+    # Trigger a check
+    client.get(url_for("api_watch_checknow"), follow_redirects=True)
+
+    # Give the thread time to pick it up
+    time.sleep(sleep_time_for_fetch_thread)
+
+    # It should report nothing found (no new 'unviewed' class)
+    res = client.get(url_for("index"))
+    assert b'unviewed' in res.data

changedetectionio/tests/util.py

@@ -37,6 +37,16 @@ def set_modified_response():
 
 def live_server_setup(live_server):
 
+    @live_server.app.route('/test-binary-endpoint')
+    def test_binaryfile_endpoint():
+
+        from flask import make_response
+
+        # Tried using a global var here but didn't seem to work, so reading from a file instead.
+        with open("test-datastore/test.bin", "rb") as f:
+            resp = make_response(f.read())
+            resp.headers['Content-Type'] = 'image/jpeg'
+            return resp
 
     @live_server.app.route('/test-endpoint')
     def test_endpoint():

changedetectionio/update_worker.py

@@ -42,7 +42,6 @@ class update_worker(threading.Thread):
                     now = time.time()
 
                     try:
-
                         changed_detected, update_obj, contents = update_handler.run(uuid)
 
                         # Re #342
@@ -135,8 +134,8 @@ class update_worker(threading.Thread):
 
                         except Exception as e:
                             # Catch everything possible here, so that if a worker crashes, we don't lose it until restart!
-                            print("!!!! Exception in update_worker !!!\n", e)
-
+                            self.app.logger.error("Exception reached processing watch UUID: %s - %s", uuid, str(e))
+                            self.datastore.update_watch(uuid=uuid, update_obj={'last_error': str(e)})
                     finally:
                         # Always record that we atleast tried
                         self.datastore.update_watch(uuid=uuid, update_obj={'fetch_time': round(time.time() - now, 3),