From a18fa8d28834f95e011d6cefac4cbdad9a8ac26c Mon Sep 17 00:00:00 2001
From: dgtlmoon <dgtlmoon@gmail.com>
Date: Tue, 31 Oct 2023 18:52:39 +0100
Subject: [PATCH] Ability to import wachete XLSX

---
 changedetectionio/__init__.py           | 15 +++++--
 changedetectionio/forms.py              |  2 +
 changedetectionio/importer.py           | 55 +++++++++++++++++++++++++
 changedetectionio/templates/import.html |  6 ++-
 requirements.txt                        |  2 +-
 5 files changed, 75 insertions(+), 5 deletions(-)

diff --git a/changedetectionio/__init__.py b/changedetectionio/__init__.py
index 850e9966f..f08732e90 100644
--- a/changedetectionio/__init__.py
+++ b/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 +846,19 @@ 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}))
 
+            # Wachete import
+            from .forms import importForm
+            form = importForm()
+            if request.files:
+                file = request.files['wachete_export']
+                from .importer import import_wachete_xlsx
+                w_importer = import_wachete_xlsx()
+                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),
diff --git a/changedetectionio/forms.py b/changedetectionio/forms.py
index 4177ed802..d393fdda2 100644
--- a/changedetectionio/forms.py
+++ b/changedetectionio/forms.py
@@ -15,6 +15,7 @@ from wtforms import (
     validators,
     widgets
 )
+from flask_wtf.file import FileField, FileAllowed
 from wtforms.fields import FieldList
 from wtforms.validators import ValidationError
 
@@ -398,6 +399,7 @@ class importForm(Form):
     from . import processors
     processor = RadioField(u'Processor', choices=processors.available_processors(), default="text_json_diff")
     urls = TextAreaField('URLs')
+    wachete_export = FileField(validators=[FileAllowed(['xlsx'], 'Must be .xlsx file!')])
 
 class SingleBrowserStep(Form):
 
diff --git a/changedetectionio/importer.py b/changedetectionio/importer.py
index f046c8952..4adae9d76 100644
--- a/changedetectionio/importer.py
+++ b/changedetectionio/importer.py
@@ -132,3 +132,58 @@ 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_wachete_xlsx(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('title'):
+                extras['title'] = [data.get('title').strip()]
+
+            new_uuid = datastore.add_watch(url=data['url'].strip(),
+                                           extras=extras,
+                                           tag=data.get('folder'),
+                                           write_to_disk_now=False)
+            row += 1
+            i = 1
+
+            if new_uuid:
+                # Straight into the queue.
+                self.new_uuids.append(new_uuid)
+                good += 1
+
+        flash(
+            "{} Imported from Wachete xlsx.io in {:.2f}s, {} Skipped.".format(len(self.new_uuids), time.time() - now, 0))
diff --git a/changedetectionio/templates/import.html b/changedetectionio/templates/import.html
index 9824ccf56..ab02e763b 100644
--- a/changedetectionio/templates/import.html
+++ b/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="#wachete">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 pure-form-aligned" 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 +80,9 @@
 " rows="25">{{ original_distill_json }}</textarea>
 
             </div>
+            <div class="tab-pane-inner" id="wachete">
+                {{ render_field(form.wachete_export, class="processor") }}
+            </div>
             <button type="submit" class="pure-button pure-input-1-2 pure-button-primary">Import</button>
         </form>
 
diff --git a/requirements.txt b/requirements.txt
index 8d0500665..1ffe38c0b 100644
--- a/requirements.txt
+++ b/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"