2017-07-23 19:56:51 +02:00
|
|
|
{% extends "main.html" %}
|
2017-09-07 13:40:42 +02:00
|
|
|
{% block title %}Scrape{% endblock %}
|
2017-07-23 19:56:51 +02:00
|
|
|
|
|
|
|
{% block content %}
|
|
|
|
<div class="container">
|
2019-06-28 11:59:34 +02:00
|
|
|
<center>
|
|
|
|
<img src="{{ url_for('static', filename='lookyloo.jpeg') }}"
|
2020-02-10 01:55:03 +01:00
|
|
|
alt="Lookyloo" width="400">
|
2019-06-28 11:59:34 +02:00
|
|
|
</center>
|
|
|
|
</br>
|
2020-08-25 15:51:08 +02:00
|
|
|
<form role="form" action="{{ url_for('scrape_web') }}" method=post enctype=multipart/form-data>
|
2020-05-22 17:36:22 +02:00
|
|
|
<div class="form-group row">
|
|
|
|
<div class="col-sm-10">
|
|
|
|
<div class="form-check">
|
|
|
|
<input class="form-check-input"type="checkbox" name="listing" checked="true"></input>
|
|
|
|
<label for="listing" class="form-check-label">Display results on public page</label>
|
|
|
|
</div>
|
|
|
|
</div>
|
|
|
|
</div>
|
2019-06-28 11:59:34 +02:00
|
|
|
<div class="form-group row">
|
|
|
|
<label for="url" class="col-sm-2 col-form-label">URL:</label>
|
|
|
|
<div class="col-sm-10">
|
2020-07-08 00:37:29 +02:00
|
|
|
<input type="text" class="form-control" name="url" id=url placeholder="URL to capture" required>
|
2019-06-27 17:47:21 +02:00
|
|
|
</div>
|
2019-06-28 11:59:34 +02:00
|
|
|
</div>
|
2019-06-27 17:47:21 +02:00
|
|
|
|
2019-06-28 11:59:34 +02:00
|
|
|
<div class="form-group row">
|
|
|
|
<label for="depth" class="col-sm-2 col-form-label">Link Depth:</label>
|
|
|
|
<div class="col-sm-1">
|
|
|
|
<select class="form-control" name="depth" id=depth>
|
|
|
|
<option value="1">1</option>
|
|
|
|
<option value="2">2</option>
|
|
|
|
<option value="3">3</option>
|
|
|
|
<option value="4">4</option>
|
|
|
|
<option value="5">5</option>
|
|
|
|
<option value="6">6</option>
|
|
|
|
<option value="7">7</option>
|
|
|
|
<option value="8">8</option>
|
|
|
|
<option value="9">9</option>
|
|
|
|
</select>
|
2017-07-23 19:56:51 +02:00
|
|
|
</div>
|
2019-06-28 11:59:34 +02:00
|
|
|
</div>
|
2019-06-27 17:47:21 +02:00
|
|
|
|
2020-07-08 00:37:29 +02:00
|
|
|
<div class="form-group row">
|
2020-07-17 18:39:50 +02:00
|
|
|
<label for="referer" class="col-sm-2 col-form-label">Referer (Optional):</label>
|
2020-07-08 00:37:29 +02:00
|
|
|
<div class="col-sm-10">
|
2020-07-17 18:39:50 +02:00
|
|
|
<input type="text" class="form-control" name="referer" id=referer placeholder="Pass referer of the URL">
|
2020-07-08 00:37:29 +02:00
|
|
|
</div>
|
|
|
|
</div>
|
|
|
|
|
2020-01-24 10:17:41 +01:00
|
|
|
<div class="form-group row">
|
|
|
|
<label for="cookies" class="col-sm-6">Cookies (JSON export from the Firefox plugin Cookie Quick Manager)</label>
|
|
|
|
<div class="col-sm-4">
|
2020-01-24 11:25:53 +01:00
|
|
|
<input type="file" class="form-control-file" id="cookies" name="cookies">
|
2020-01-24 10:17:41 +01:00
|
|
|
</div>
|
|
|
|
</div>
|
|
|
|
|
2019-06-28 11:59:34 +02:00
|
|
|
<div class="dropdown-divider"></div>
|
|
|
|
<p><u>Browser Configuration</u></p>
|
|
|
|
<div class="form-group row">
|
|
|
|
<label for="os" class="col-sm-2 col-form-label">Operating System:</label>
|
|
|
|
<div class="col-sm-10">
|
|
|
|
<select class="form-control" name="os" id="os">
|
|
|
|
{% for os in user_agents.keys() %}
|
|
|
|
<option value="{{ os }}">{{ os }}</option>
|
2019-06-27 17:47:21 +02:00
|
|
|
{% endfor%}
|
2019-06-28 11:59:34 +02:00
|
|
|
</select>
|
2019-03-29 20:11:44 +01:00
|
|
|
</div>
|
2019-06-28 11:59:34 +02:00
|
|
|
</div>
|
|
|
|
|
|
|
|
{% for os, browsers in user_agents.items() %}
|
2019-07-05 15:44:52 +02:00
|
|
|
<div id="{{os.replace(' ', '_')}}" class="style-sub-1 form-group row" {% if not loop.first %}style="display: none;"{%endif%}>
|
2019-06-28 11:59:34 +02:00
|
|
|
<label for="browser" class="col-sm-2 col-form-label">Browser Type:</label>
|
|
|
|
<div class="col-sm-10">
|
2019-07-05 15:44:52 +02:00
|
|
|
<!-- Disable all the selects, unless it is the first one in the OS list-->
|
|
|
|
<select class="form-control" name="browser" {% if not loop.first %}disabled{%endif%}>
|
2019-06-28 11:59:34 +02:00
|
|
|
{% for browser in browsers.keys()%}
|
|
|
|
<option value="{{ browser }}">{{ browser }}</option>
|
|
|
|
{% endfor%}
|
|
|
|
</select>
|
|
|
|
</div>
|
|
|
|
</div>
|
2019-06-27 17:47:21 +02:00
|
|
|
{% set outer_loop = loop %}
|
|
|
|
{% for browser, user_agents in browsers.items()%}
|
2019-06-28 11:59:34 +02:00
|
|
|
<div id="{{os.replace(' ', '_')}}_{{browser.replace(' ', '_')}}" class="style-sub-2 form-group row" {% if not loop.first or not outer_loop.first %} style="display: none;"{%endif%}>
|
|
|
|
<label for="user_agent" class="col-sm-2 col-form-label">User Agent:</label>
|
|
|
|
<div class="col-sm-10">
|
2019-07-05 15:44:52 +02:00
|
|
|
<!-- Disable all the selects, unless it is the first one in the OS list and in the browser list-->
|
|
|
|
<select class="form-control" name="user_agent" {% if not outer_loop.first or not loop.first %}disabled{%endif%}>
|
2019-06-27 17:47:21 +02:00
|
|
|
{% for user_agent in user_agents %}
|
2019-07-05 15:44:52 +02:00
|
|
|
<option value="{{ user_agent }}">{{ user_agent }}</option>
|
2019-03-29 20:11:44 +01:00
|
|
|
{% endfor%}
|
2019-06-28 11:59:34 +02:00
|
|
|
</select>
|
2019-06-27 17:47:21 +02:00
|
|
|
</div>
|
2019-06-28 11:59:34 +02:00
|
|
|
</div>
|
2019-06-27 17:47:21 +02:00
|
|
|
{% endfor%}
|
2019-06-28 11:59:34 +02:00
|
|
|
{% endfor%}
|
|
|
|
<div class="dropdown-divider"></div>
|
2020-07-17 18:39:50 +02:00
|
|
|
<center>
|
|
|
|
<button type="submit" class="btn btn-primary" id="btn-looking">Start looking!</button>
|
2020-08-07 17:06:25 +02:00
|
|
|
<div id="long" style="display:none">Please wait, it can take a while on big websites...</div>
|
|
|
|
<div id="too-long" style="display:none">... No, really, I mean it.</div>
|
|
|
|
<div id="way-too-long" style="display:none">Told ya!</div>
|
2020-07-17 18:39:50 +02:00
|
|
|
</center>
|
2019-06-28 11:59:34 +02:00
|
|
|
</form>
|
2017-07-23 19:56:51 +02:00
|
|
|
</div>
|
|
|
|
{% endblock %}
|
2019-03-29 20:11:44 +01:00
|
|
|
|
|
|
|
{% block scripts %}
|
2019-06-28 11:59:34 +02:00
|
|
|
{{ super() }}
|
2019-07-05 15:44:52 +02:00
|
|
|
<script src='{{ url_for('static', filename='scrape.js') }}'></script>
|
2019-03-29 20:11:44 +01:00
|
|
|
{% endblock %}
|