From 361264aef739a0c69708534b2741bd802415e8de Mon Sep 17 00:00:00 2001
From: Georg Petz <georg.petz@onb.ac.at>
Date: Wed, 7 Apr 2021 12:06:25 +0200
Subject: [PATCH] add zip

---
 txtDownloader.ipynb | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/txtDownloader.ipynb b/txtDownloader.ipynb
index 9d60c5f..865ad53 100644
--- a/txtDownloader.ipynb
+++ b/txtDownloader.ipynb
@@ -2,16 +2,17 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 4,
    "metadata": {},
    "outputs": [],
    "source": [
-    "import urllib.request, json, shutil"
+    "import urllib.request, json, shutil, os\n",
+    "from zipfile import ZipFile"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -20,7 +21,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 6,
    "metadata": {},
    "outputs": [
     {
@@ -51,6 +52,7 @@
     "data = json.loads(response.read())\n",
     "\n",
     "i=0\n",
+    "zipObj = ZipFile('downloadTXT.zip', 'w')\n",
     "\n",
     "for page in data[\"sequences\"][0][\"canvases\"]:\n",
     "    txt_url = page[\"otherContent\"][0][\"resources\"][0][\"resource\"][\"@id\"]\n",
@@ -58,7 +60,10 @@
     "    print(txt_url)\n",
     "    output_file = str(i)+\".txt\"\n",
     "    with urllib.request.urlopen(txt_url) as response, open(output_file, 'wb') as out_file:\n",
-    "        shutil.copyfileobj(response, out_file)"
+    "        shutil.copyfileobj(response, out_file)\n",
+    "    zipObj.write(os.path.abspath(output_file))\n",
+    "    \n",
+    "zipObj.close()"
    ]
   },
   {
-- 
GitLab