data: add notebook for creating dataset splits

CreateDatasetSplits.ipynb

@@ -0,0 +1,232 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "42ac3dd1-7154-40ec-ae54-38e4389c5ea8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import random\n",
+    "import requests\n",
+    "import zipfile\n",
+    "\n",
+    "from io import BytesIO"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "1ba83fb8-8c38-427e-83c1-68da2b5b4bbd",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Downloading dataset from https://coltekin.github.io/offensive-turkish/offenseval2020-turkish.zip\n",
+      "Extracting files to './'...\n",
+      "Extracted files: ['offenseval2020-turkish/', 'offenseval2020-turkish/offenseval-tr-training-v1/', 'offenseval2020-turkish/offenseval-tr-training-v1/offenseval-annotation.txt', 'offenseval2020-turkish/offenseval-tr-training-v1/offenseval-tr-training-v1.tsv', 'offenseval2020-turkish/offenseval-tr-training-v1/readme-trainingset-tr.txt', 'offenseval2020-turkish/offenseval-tr-testset-v1/', 'offenseval2020-turkish/offenseval-tr-testset-v1/offenseval-tr-testset-v1.tsv', 'offenseval2020-turkish/offenseval-tr-testset-v1/offenseval-tr-labela-v1.tsv', 'offenseval2020-turkish/README.txt']\n"
+     ]
+    }
+   ],
+   "source": [
+    "def download_and_extract_zip(url, extract_to=\"./\"):\n",
+    "    try:\n",
+    "        print(f\"Downloading dataset from {url}\")\n",
+    "        response = requests.get(url)\n",
+    "        response.raise_for_status()\n",
+    "\n",
+    "        with zipfile.ZipFile(BytesIO(response.content)) as z:\n",
+    "            print(f\"Extracting files to '{extract_to}'...\")\n",
+    "            z.extractall(extract_to)\n",
+    "            extracted_files = z.namelist()\n",
+    "            print(f\"Extracted files: {extracted_files}\")\n",
+    "    except Exception as e:\n",
+    "        print(f\"An error occurred: {e}\")\n",
+    "\n",
+    "url = \"https://coltekin.github.io/offensive-turkish/offenseval2020-turkish.zip\"  # Replace with the actual URL\n",
+    "download_and_extract_zip(url, \"./\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "b74682a7-ccf8-44ad-98a0-73636c35e10e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "original_train_file = \"./offenseval2020-turkish/offenseval-tr-training-v1/offenseval-tr-training-v1.tsv\"\n",
+    "original_test_file = \"./offenseval2020-turkish/offenseval-tr-testset-v1/offenseval-tr-testset-v1.tsv\"\n",
+    "orginal_label_test_file = \"./offenseval2020-turkish/offenseval-tr-testset-v1/offenseval-tr-labela-v1.tsv\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "af3ba702-1341-4e70-8c74-87078eeaddf1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_instances(filename: str):\n",
+    "    instances = []\n",
+    "    with open(filename, \"rt\") as f_p:\n",
+    "        for line in f_p:\n",
+    "            line = line.strip()\n",
+    "    \n",
+    "            if not line:\n",
+    "                continue\n",
+    "    \n",
+    "            if line.startswith(\"id\"):\n",
+    "                continue\n",
+    "    \n",
+    "            _, tweet, label = line.split(\"\\t\")\n",
+    "\n",
+    "            instances.append([label, tweet])\n",
+    "\n",
+    "    print(f\"Found {len(instances)} training instances.\")\n",
+    "\n",
+    "    return instances"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "a397f04d-354c-4d97-9f93-794929c5e51d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_test_instances(filename: str, label_filename: str):\n",
+    "    # E.g. 41993,NOT is mapped to \"41993\" -> \"NOT\"\n",
+    "    id_label_mapping = {}\n",
+    "    with open(label_filename, \"rt\") as f_p:\n",
+    "        for line in f_p:\n",
+    "            line = line.strip()\n",
+    "\n",
+    "            if not line:\n",
+    "                continue\n",
+    "\n",
+    "            id_, label = line.split(\",\")\n",
+    "\n",
+    "            id_label_mapping[id_] = label\n",
+    "\n",
+    "    print(f\"Found {len(id_label_mapping)} labelled test instances\")\n",
+    "\n",
+    "    instances = []\n",
+    "    \n",
+    "    with open(filename, \"rt\") as f_p:\n",
+    "        for line in f_p:\n",
+    "            line = line.strip()\n",
+    "    \n",
+    "            if not line:\n",
+    "                continue\n",
+    "    \n",
+    "            if line.startswith(\"id\"):\n",
+    "                continue\n",
+    "    \n",
+    "            id_, tweet = line.split(\"\\t\")\n",
+    "\n",
+    "            label = id_label_mapping[id_]\n",
+    "\n",
+    "            instances.append([label, tweet])\n",
+    "    return instances\n",
+    "\n",
+    "    assert len(id_label_mapping) == len(instances)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "06508ca8-649b-44ea-a8c6-09c2c2b434f4",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 31756 training instances.\n",
+      "Found 3528 labelled test instances\n"
+     ]
+    }
+   ],
+   "source": [
+    "original_train_instances = get_instances(original_train_file)\n",
+    "original_test_instances = get_test_instances(original_test_file, orginal_label_test_file)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "cd4a7942-42c9-4bdb-9079-c6039ff908c4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Shuffling is done in-place\n",
+    "random.seed(83607)\n",
+    "random.shuffle(original_train_instances)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "6ab61fde-e534-4ffe-9302-f80581e503eb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_instances = original_train_instances[:30_000]\n",
+    "dev_instances = original_train_instances[30_000:]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "84f9c044-7906-4c04-9154-70e2b8d55982",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def write_instances(instances: str, split_name: str):\n",
+    "    with open(f\"{split_name}.txt\", \"wt\") as f_out:\n",
+    "        for instance in instances:\n",
+    "            label, tweet = instance\n",
+    "\n",
+    "            # We stick to Flair format for classification tasks, which is basically FastText inspired ;)\n",
+    "            new_label = \"__label__\" + label\n",
+    "            f_out.write(f\"{new_label} {tweet}\\n\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "0bf06e96-2b25-46ed-8a7e-0672e7aa6af8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "write_instances(train_instances, \"train\")\n",
+    "write_instances(dev_instances, \"dev\")\n",
+    "write_instances(original_test_instances, \"test\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}