Ability to Import from Wachete XLSX (or any XLSX) - Wachete alternative made easy (#1921)

changedetectionio/__init__.py

@@ -822,6 +822,7 @@ def changedetection_app(config=None, datastore_o=None):
         from . import forms
 
         if request.method == 'POST':
+
             from .importer import import_url_list, import_distill_io_json
 
             # URL List import
@@ -845,11 +846,32 @@ def changedetection_app(config=None, datastore_o=None):
                 for uuid in d_importer.new_uuids:
                     update_q.put(queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid, 'skip_when_checksum_same': True}))
 
+            # XLSX importer
+            if request.files and request.files.get('xlsx_file'):
+                file = request.files['xlsx_file']
+                from .importer import import_xlsx_wachete, import_xlsx_custom
+
+                if request.values.get('file_mapping') == 'wachete':
+                    w_importer = import_xlsx_wachete()
+                    w_importer.run(data=file, flash=flash, datastore=datastore)
+                else:
+                    w_importer = import_xlsx_custom()
+                    # Building mapping of col # to col # type
+                    map = {}
+                    for i in range(10):
+                        c = request.values.get(f"custom_xlsx[col_{i}]")
+                        v = request.values.get(f"custom_xlsx[col_type_{i}]")
+                        if c and v:
+                            map[int(c)] = v
+
+                    w_importer.import_profile = map
+                    w_importer.run(data=file, flash=flash, datastore=datastore)
+
+                for uuid in w_importer.new_uuids:
+                    update_q.put(queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid, 'skip_when_checksum_same': True}))
 
-        form = forms.importForm(formdata=request.form if request.method == 'POST' else None,
-#                               data=default,
-                               )
         # Could be some remaining, or we could be on GET
+        form = forms.importForm(formdata=request.form if request.method == 'POST' else None)
         output = render_template("import.html",
                                  form=form,
                                  import_url_list_remaining="\n".join(remaining_urls),

changedetectionio/forms.py

@@ -15,9 +15,14 @@ from wtforms import (
     validators,
     widgets
 )
+from flask_wtf.file import FileField, FileAllowed
 from wtforms.fields import FieldList
+
 from wtforms.validators import ValidationError
 
+from validators.url import url as url_validator
+
+
 # default
 # each select <option data-enabled="enabled-0-0"
 from changedetectionio.blueprint.browser_steps.browser_steps import browser_step_ui_config
@@ -41,7 +46,7 @@ valid_method = {
 }
 
 default_method = 'GET'
-
+allow_simplehost = not strtobool(os.getenv('BLOCK_SIMPLEHOSTS', 'False'))
 
 class StringListField(StringField):
     widget = widgets.TextArea()
@@ -261,19 +266,23 @@ class validateURL(object):
         self.message = message
 
     def __call__(self, form, field):
-        import validators
-        # If hosts that only contain alphanumerics are allowed ("localhost" for example)
-        allow_simplehost = not strtobool(os.getenv('BLOCK_SIMPLEHOSTS', 'False'))
-        try:
-            validators.url(field.data.strip(), simple_host=allow_simplehost)
-        except validators.ValidationFailure:
-            message = field.gettext('\'%s\' is not a valid URL.' % (field.data.strip()))
-            raise ValidationError(message)
-
-        from .model.Watch import is_safe_url
-        if not is_safe_url(field.data):
-            raise ValidationError('Watch protocol is not permitted by SAFE_PROTOCOL_REGEX')
-
+        # This should raise a ValidationError() or not
+        validate_url(field.data)
+
+def validate_url(test_url):
+    # If hosts that only contain alphanumerics are allowed ("localhost" for example)
+    try:
+        url_validator(test_url, simple_host=allow_simplehost)
+    except validators.ValidationError:
+        #@todo check for xss
+        message = f"'{test_url}' is not a valid URL."
+        # This should be wtforms.validators.
+        raise ValidationError(message)
+
+    from .model.Watch import is_safe_url
+    if not is_safe_url(test_url):
+        # This should be wtforms.validators.
+        raise ValidationError('Watch protocol is not permitted by SAFE_PROTOCOL_REGEX or incorrect URL format')
 
 class ValidateListRegex(object):
     """
@@ -398,6 +407,9 @@ class importForm(Form):
     from . import processors
     processor = RadioField(u'Processor', choices=processors.available_processors(), default="text_json_diff")
     urls = TextAreaField('URLs')
+    xlsx_file = FileField('Upload .xlsx file', validators=[FileAllowed(['xlsx'], 'Must be .xlsx file!')])
+    file_mapping = SelectField('File mapping', [validators.DataRequired()], choices={('wachete', 'Wachete mapping'), ('custom','Custom mapping')})
+
 
 class SingleBrowserStep(Form):
 

changedetectionio/importer.py

@@ -1,6 +1,9 @@
 from abc import ABC, abstractmethod
 import time
 import validators
+from wtforms import ValidationError
+
+from changedetectionio.forms import validate_url
 
 
 class Importer():
@@ -12,6 +15,7 @@ class Importer():
         self.new_uuids = []
         self.good = 0
         self.remaining_data = []
+        self.import_profile = None
 
     @abstractmethod
     def run(self,
@@ -132,3 +136,145 @@ class import_distill_io_json(Importer):
                     good += 1
 
         flash("{} Imported from Distill.io in {:.2f}s, {} Skipped.".format(len(self.new_uuids), time.time() - now, len(self.remaining_data)))
+
+class import_xlsx_wachete(Importer):
+
+    def run(self,
+            data,
+            flash,
+            datastore,
+            ):
+        good = 0
+        now = time.time()
+        self.new_uuids = []
+
+        from openpyxl import load_workbook
+
+        try:
+            wb = load_workbook(data)
+        except Exception as e:
+            #@todo correct except
+            flash("Unable to read export XLSX file, something wrong with the file?", 'error')
+            return
+
+        sheet_obj = wb.active
+
+        i = 1
+        row = 2
+        while sheet_obj.cell(row=row, column=1).value:
+            data = {}
+            while sheet_obj.cell(row=row, column=i).value:
+                column_title = sheet_obj.cell(row=1, column=i).value.strip().lower()
+                column_row_value = sheet_obj.cell(row=row, column=i).value
+                data[column_title] = column_row_value
+
+                i += 1
+
+            extras = {}
+            if data.get('xpath'):
+                #@todo split by || ?
+                extras['include_filters'] = [data.get('xpath')]
+            if data.get('name'):
+                extras['title'] = [data.get('name').strip()]
+            if data.get('interval (min)'):
+                minutes = int(data.get('interval (min)'))
+                hours, minutes = divmod(minutes, 60)
+                days, hours = divmod(hours, 24)
+                weeks, days = divmod(days, 7)
+                extras['time_between_check'] = {'weeks': weeks, 'days': days, 'hours': hours, 'minutes': minutes, 'seconds': 0}
+
+
+            # At minimum a URL is required.
+            if data.get('url'):
+                try:
+                    validate_url(data.get('url'))
+                except ValidationError as e:
+                    print(">> import URL error", data.get('url'), str(e))
+                    # Don't bother processing anything else on this row
+                    continue
+
+                new_uuid = datastore.add_watch(url=data['url'].strip(),
+                                               extras=extras,
+                                               tag=data.get('folder'),
+                                               write_to_disk_now=False)
+                if new_uuid:
+                    # Straight into the queue.
+                    self.new_uuids.append(new_uuid)
+                    good += 1
+
+            row += 1
+            i = 1
+
+
+        flash(
+            "{} imported from Wachete .xlsx in {:.2f}s".format(len(self.new_uuids), time.time() - now))
+
+class import_xlsx_custom(Importer):
+
+    def run(self,
+            data,
+            flash,
+            datastore,
+            ):
+        good = 0
+        now = time.time()
+        self.new_uuids = []
+
+        from openpyxl import load_workbook
+
+        try:
+            wb = load_workbook(data)
+        except Exception as e:
+            #@todo correct except
+            flash("Unable to read export XLSX file, something wrong with the file?", 'error')
+            return
+
+        # @todo cehck atleast 2 rows, same in other method
+
+        sheet_obj = wb.active
+        from .forms import validate_url
+        row = 2
+        while sheet_obj.cell(row=row, column=1).value:
+            url = None
+            tags = None
+            extras = {}
+            for col_i, cell_map in self.import_profile.items():
+                cell_val = sheet_obj.cell(row=row, column=col_i).value
+                if cell_map == 'url':
+                    url = cell_val.strip()
+                    try:
+                        validate_url(url)
+                    except ValidationError as e:
+                        print (">> Import URL error",url, str(e))
+                        # Don't bother processing anything else on this row
+                        url = None
+                        break
+
+                elif cell_map == 'tag':
+                    tags = cell_val.strip()
+                elif cell_map == 'include_filters':
+                    # @todo validate?
+                    extras['include_filters'] = [cell_val.strip()]
+                elif cell_map == 'interval_minutes':
+                    hours, minutes = divmod(int(cell_val), 60)
+                    days, hours = divmod(hours, 24)
+                    weeks, days = divmod(days, 7)
+                    extras['time_between_check'] = {'weeks': weeks, 'days': days, 'hours': hours, 'minutes': minutes, 'seconds': 0}
+                else:
+                    extras[cell_map] = cell_val.strip()
+
+            # At minimum a URL is required.
+            if url:
+                new_uuid = datastore.add_watch(url=url,
+                                               extras=extras,
+                                               tag=tags,
+                                               write_to_disk_now=False)
+                if new_uuid:
+                    # Straight into the queue.
+                    self.new_uuids.append(new_uuid)
+                    good += 1
+
+            row += 1
+
+        flash(
+            "{} imported from custom .xlsx in {:.2f}s".format(len(self.new_uuids), time.time() - now))

changedetectionio/store.py

@@ -360,6 +360,8 @@ class ChangeDetectionStore:
         if write_to_disk_now:
             self.sync_to_json()
 
+        print("added ", url)
+
         return new_uuid
 
     def visualselector_data_is_ready(self, watch_uuid):

changedetectionio/templates/import.html

@@ -8,11 +8,12 @@
         <ul>
             <li class="tab" id=""><a href="#url-list">URL List</a></li>
             <li class="tab"><a href="#distill-io">Distill.io</a></li>
+            <li class="tab"><a href="#xlsx">.XLSX &amp; Wachete</a></li>
         </ul>
     </div>
 
     <div class="box-wrap inner">
-        <form class="pure-form pure-form-aligned" action="{{url_for('import_page')}}" method="POST">
+        <form class="pure-form" action="{{url_for('import_page')}}" method="POST" enctype="multipart/form-data">
             <input type="hidden" name="csrf_token" value="{{ csrf_token() }}">
             <div class="tab-pane-inner" id="url-list">
                     <legend>
@@ -79,6 +80,42 @@
 " rows="25">{{ original_distill_json }}</textarea>
 
             </div>
+            <div class="tab-pane-inner" id="xlsx">
+            <fieldset>
+                <div class="pure-control-group">
+                {{ render_field(form.xlsx_file, class="processor") }}
+                </div>
+                <div class="pure-control-group">
+                    {{ render_field(form.file_mapping, class="processor") }}
+                </div>
+            </fieldset>
+                <div class="pure-control-group">
+                <span class="pure-form-message-inline">
+                    Table of custom column and data types mapping for the <strong>Custom mapping</strong> File mapping type.
+                </span>
+                    <table style="border: 1px solid #aaa; padding: 0.5rem; border-radius: 4px;">
+                        <tr>
+                            <td><strong>Column #</strong></td>
+                            {% for n in range(4) %}
+                                <td><input type="number" name="custom_xlsx[col_{{n}}]" style="width: 4rem;" min="1"></td>
+                            {%  endfor %}
+                        </tr>
+                        <tr>
+                            <td><strong>Type</strong></td>
+                            {% for n in range(4) %}
+                                <td><select name="custom_xlsx[col_type_{{n}}]">
+                                    <option value="" style="color: #aaa"> -- none --</option>
+                                    <option value="url">URL</option>
+                                    <option value="title">Title</option>
+                                    <option value="include_filter">CSS/xPath filter</option>
+                                    <option value="tag">Group / Tag name(s)</option>
+                                    <option value="interval_minutes">Recheck time (minutes)</option>
+                                </select></td>
+                            {%  endfor %}
+                        </tr>
+                    </table>
+                </div>
+            </div>
             <button type="submit" class="pure-button pure-input-1-2 pure-button-primary">Import</button>
         </form>
 

changedetectionio/tests/test_import.py

@@ -1,16 +1,19 @@
 #!/usr/bin/python3
-
+import io
+import os
 import time
 
 from flask import url_for
 
-from .util import live_server_setup
+from .util import live_server_setup, wait_for_all_checks
+
+
 def test_setup(client, live_server):
     live_server_setup(live_server)
 
 def test_import(client, live_server):
     # Give the endpoint time to spin up
-    time.sleep(1)
+    wait_for_all_checks(client)
 
     res = client.post(
         url_for("import_page"),
@@ -119,3 +122,82 @@ def test_import_distillio(client, live_server):
     res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
     # Clear flask alerts
     res = client.get(url_for("index"))
+
+def test_import_custom_xlsx(client, live_server):
+    """Test can upload a excel spreadsheet and the watches are created correctly"""
+
+    #live_server_setup(live_server)
+    dirname = os.path.dirname(__file__)
+    filename = os.path.join(dirname, 'import/spreadsheet.xlsx')
+    with open(filename, 'rb') as f:
+
+        data= {
+            'file_mapping': 'custom',
+            'custom_xlsx[col_0]': '1',
+            'custom_xlsx[col_1]': '3',
+            'custom_xlsx[col_2]': '5',
+            'custom_xlsx[col_3]': '4',
+            'custom_xlsx[col_type_0]': 'title',
+            'custom_xlsx[col_type_1]': 'url',
+            'custom_xlsx[col_type_2]': 'include_filters',
+            'custom_xlsx[col_type_3]': 'interval_minutes',
+            'xlsx_file': (io.BytesIO(f.read()), 'spreadsheet.xlsx')
+        }
+
+    res = client.post(
+        url_for("import_page"),
+        data=data,
+        follow_redirects=True,
+    )
+
+    assert b'2 imported from custom .xlsx' in res.data
+
+    res = client.get(
+        url_for("index")
+    )
+
+
+    assert b'Somesite results ABC' in res.data
+    assert b'City news results' in res.data
+
+    # Just find one to check over
+    for uuid, watch in live_server.app.config['DATASTORE'].data['watching'].items():
+        if watch.get('title') == 'Somesite results ABC':
+            filters = watch.get('include_filters')
+            assert filters[0] == '/html[1]/body[1]/div[4]/div[1]/div[1]/div[1]||//*[@id=\'content\']/div[3]/div[1]/div[1]||//*[@id=\'content\']/div[1]'
+            assert watch.get('time_between_check') == {'weeks': 0, 'days': 1, 'hours': 6, 'minutes': 24, 'seconds': 0}
+
+def test_import_watchete_xlsx(client, live_server):
+    """Test can upload a excel spreadsheet and the watches are created correctly"""
+
+    #live_server_setup(live_server)
+    dirname = os.path.dirname(__file__)
+    filename = os.path.join(dirname, 'import/spreadsheet.xlsx')
+    with open(filename, 'rb') as f:
+
+        data= {
+            'file_mapping': 'wachete',
+            'xlsx_file': (io.BytesIO(f.read()), 'spreadsheet.xlsx')
+        }
+
+    res = client.post(
+        url_for("import_page"),
+        data=data,
+        follow_redirects=True,
+    )
+
+    assert b'2 imported from Wachete .xlsx' in res.data
+
+    res = client.get(
+        url_for("index")
+    )
+
+    assert b'Somesite results ABC' in res.data
+    assert b'City news results' in res.data
+
+    # Just find one to check over
+    for uuid, watch in live_server.app.config['DATASTORE'].data['watching'].items():
+        if watch.get('title') == 'Somesite results ABC':
+            filters = watch.get('include_filters')
+            assert filters[0] == '/html[1]/body[1]/div[4]/div[1]/div[1]/div[1]||//*[@id=\'content\']/div[3]/div[1]/div[1]||//*[@id=\'content\']/div[1]'
+            assert watch.get('time_between_check') == {'weeks': 0, 'days': 1, 'hours': 6, 'minutes': 24, 'seconds': 0}

requirements.txt

@@ -59,7 +59,7 @@ werkzeug~=2.0.0
 # Templating, so far just in the URLs but in the future can be for the notifications also
 jinja2~=3.1
 jinja2-time
-
+openpyxl
 # https://peps.python.org/pep-0508/#environment-markers
 # https://github.com/dgtlmoon/changedetection.io/pull/1009
 jq~=1.3; python_version >= "3.8" and sys_platform == "darwin"