updates in oracc link; reading project that failed

project_notebook.ipynb

@@ -21,6 +21,7 @@
     "import re\n",
     "import logging\n",
     "import datetime\n",
+    "import os\n",
     "logging.basicConfig(\n",
     "    level=logging.INFO, filename=f'{datetime.datetime.now().strftime(\"%Y-%m-%d_%H-%M-%S\")}.log', filemode='w',\n",
     "    format='%(name)s - %(levelname)s - %(message)s', datefmt='%d-%b-%y %H:%M:%S')\n",
@@ -29,7 +30,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 28,
+   "execution_count": 29,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -56,7 +57,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 33,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -68,11 +69,14 @@
     "logging.info('Start')\n",
     "with open('project_list') as f:\n",
     "    # project = f.read().split('\\n')[2]\n",
-    "    scrap = scrapping.get_raw_english_texts_of_project('contrib/amarna')\n",
-    "    execute_batch(\n",
-    "        connection, \"insert into raw_texts values (%(id_text)s, %(project_name)s, %(raw_text)s) ON CONFLICT DO NOTHING\",\n",
-    "        scrap)\n",
-    "    conn.commit()\n"
+    "    for project in f.read().split('\\n'):\n",
+    "        # connection = connection.execute(\"insert into raw_texts values (%(id_text)s, %(project_name)s, %(raw_text)s)\", scrap)\n",
+    "        if os.path.exists(project):\n",
+    "            scrap = json.load(project)\n",
+    "            execute_batch(\n",
+    "                connection, \"insert into raw_texts values (%(id_text)s, %(project_name)s, %(raw_text)s) ON CONFLICT DO NOTHING\",\n",
+    "                scrap)\n",
+    "            conn.commit()\n"
    ]
   },
   {