From 3c9b28b704b42f2f950cae385eba2133df5a5cfd Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rapha=C3=ABl=20Vinot?= <raphael@vinot.info>
Date: Thu, 19 Jul 2018 18:18:22 +0200
Subject: [PATCH] new: Add integration with SaneJS

---
 lookyloo/__init__.py    | 37 +++++++++++++++++++++++++++++++++++++
 lookyloo/static/tree.js | 15 +++++++++++++++
 requirements.txt        |  1 +
 3 files changed, 53 insertions(+)

diff --git a/lookyloo/__init__.py b/lookyloo/__init__.py
index f5e50a23..66c309cd 100644
--- a/lookyloo/__init__.py
+++ b/lookyloo/__init__.py
@@ -19,6 +19,10 @@ import time
 from zipfile import ZipFile, ZIP_DEFLATED
 from io import BytesIO
 import base64
+import socket
+from urllib.parse import urlparse
+
+import requests
 
 app = Flask(__name__)
 
@@ -37,6 +41,27 @@ SPLASH = 'http://127.0.0.1:8050'
 
 HAR_DIR.mkdir(parents=True, exist_ok=True)
 
+SANE_JS = 'http://127.0.0.1:5007'
+
+
+def is_open(ip, port):
+    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    s.settimeout(2)
+    try:
+        s.connect((ip, int(port)))
+        s.shutdown(2)
+        return True
+    except Exception:
+        return False
+
+
+if SANE_JS:
+    parsed = urlparse(SANE_JS)
+    if is_open(parsed.hostname, parsed.port):
+        has_sane_js = True
+    else:
+        has_sane_js = False
+
 
 def cleanup_old_tmpfiles():
     for tmpfile in pathlib.Path(tempfile.gettempdir()).glob('lookyloo*'):
@@ -62,6 +87,13 @@ def load_tree(report_dir):
     return ct.to_json(), ct.start_time.isoformat(), ct.user_agent, ct.root_url
 
 
+def sane_js_query(sha512, details=False):
+    if has_sane_js:
+        r = requests.post(SANE_JS, json={"sha512": sha512, 'details': details})
+        return r.json()
+    return {}
+
+
 @app.route('/scrape', methods=['GET', 'POST'])
 def scrape():
     if request.form.get('url'):
@@ -128,6 +160,11 @@ def hostnode_details(node_uuid):
     hostnode = ct.root_hartree.get_host_node_by_uuid(node_uuid)
     urls = []
     for url in hostnode.urls:
+        if hasattr(url, 'body_hash'):
+            sane_js_r = sane_js_query(url.body_hash, details=True)
+            if sane_js_r['exists']:
+                url.add_feature('sane_js_details', sane_js_r['details'])
+                print(url.sane_js_details)
         urls.append(url.to_json())
     return json.dumps(urls)
 
diff --git a/lookyloo/static/tree.js b/lookyloo/static/tree.js
index 669537be..cf3626cd 100644
--- a/lookyloo/static/tree.js
+++ b/lookyloo/static/tree.js
@@ -267,6 +267,7 @@ function icon(icons, key, icon_path){
             .attr('x', function(d) { return d.data.total_width ? d.data.total_width + 1 : 0 })
             .attr("xlink:href", icon_path).call(getBB);
 
+
     content.filter(function(d){
             if (typeof d.data[key] === 'boolean') {
                 return false;
@@ -285,6 +286,7 @@ function icon(icons, key, icon_path){
           .attr('x', function(d) { return d.data.total_width ? d.data.total_width + 1 : 0 })
           .attr('width', function(d) { return d.to_print.toString().length + 'em'; })
           .text(function(d) { return d.to_print; }).call(getBB);
+
 };
 
 function icon_list(parent_svg, relative_x_pos, relative_y_pos) {
@@ -308,6 +310,18 @@ function icon_list(parent_svg, relative_x_pos, relative_y_pos) {
     icon(icons, 'response_cookie', "/static/cookie_received.png");
     icon(icons, 'redirect', "/static/redirect.png");
     icon(icons, 'redirect_to_nothing', "/static/cookie_in_url.png");
+
+    icons.filter(function(d){
+        if (d.data.sane_js_details) {
+            d.libname = d.data.sane_js_details[0]['libname'];
+            return d.data.sane_js_details;
+        }
+        return false;
+    }).append('text')
+      .attr('x', function(d) { return d.data.total_width ? d.data.total_width + 5 : 0 })
+      .attr('y', 15)
+      .style("font-size", "15px")
+      .text(function(d) { return 'Library name: ' + d.libname }).call(getBB);
 }
 
 function text_entry(parent_svg, relative_x_pos, relative_y_pos, onclick_callback) {
@@ -331,6 +345,7 @@ function text_entry(parent_svg, relative_x_pos, relative_y_pos, onclick_callback
               d.data.total_width = 0; // reset total_width
               to_display = d.data.name
               if (d.data.urls_count) {
+                  // Only on Hostname node.
                   to_display += ' (' + d.data.urls_count + ')';
               };
               return to_display;
diff --git a/requirements.txt b/requirements.txt
index 7796344a..73863868 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -3,6 +3,7 @@ flask
 flask-bootstrap
 ete3
 beautifulsoup4
+requests
 
 # Backend libs
 git+https://github.com/viper-framework/har2tree.git