diff --git a/l3/TM-Lab3.ipynb b/l3/TM-Lab3.ipynb
index 2f85bbd6950ea2d87c962cdd753744ec8ebd8269..6b86042744fd2ec3b8ed6c8b4fd0308d48813625 100644
--- a/l3/TM-Lab3.ipynb
+++ b/l3/TM-Lab3.ipynb
@@ -37,7 +37,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 19,
"metadata": {
"deletable": false,
"editable": false,
@@ -82,7 +82,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 20,
"metadata": {
"deletable": false,
"editable": false,
@@ -128,9 +128,103 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 21,
"metadata": {},
- "outputs": [],
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "<div>\n",
+ "<style scoped>\n",
+ " .dataframe tbody tr th:only-of-type {\n",
+ " vertical-align: middle;\n",
+ " }\n",
+ "\n",
+ " .dataframe tbody tr th {\n",
+ " vertical-align: top;\n",
+ " }\n",
+ "\n",
+ " .dataframe thead th {\n",
+ " text-align: right;\n",
+ " }\n",
+ "</style>\n",
+ "<table border=\"1\" class=\"dataframe\">\n",
+ " <thead>\n",
+ " <tr style=\"text-align: right;\">\n",
+ " <th></th>\n",
+ " <th>sentence_id</th>\n",
+ " <th>sentence</th>\n",
+ " <th>beg</th>\n",
+ " <th>end</th>\n",
+ " <th>label</th>\n",
+ " </tr>\n",
+ " </thead>\n",
+ " <tbody>\n",
+ " <tr>\n",
+ " <th>0</th>\n",
+ " <td>0000-000</td>\n",
+ " <td>EU rejects German call to boycott British lamb .</td>\n",
+ " <td>0</td>\n",
+ " <td>1</td>\n",
+ " <td>--NME--</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>1</th>\n",
+ " <td>0000-000</td>\n",
+ " <td>EU rejects German call to boycott British lamb .</td>\n",
+ " <td>2</td>\n",
+ " <td>3</td>\n",
+ " <td>Germany</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>2</th>\n",
+ " <td>0000-000</td>\n",
+ " <td>EU rejects German call to boycott British lamb .</td>\n",
+ " <td>6</td>\n",
+ " <td>7</td>\n",
+ " <td>United_Kingdom</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>3</th>\n",
+ " <td>0000-001</td>\n",
+ " <td>Peter Blackburn</td>\n",
+ " <td>0</td>\n",
+ " <td>2</td>\n",
+ " <td>--NME--</td>\n",
+ " </tr>\n",
+ " <tr>\n",
+ " <th>4</th>\n",
+ " <td>0000-002</td>\n",
+ " <td>BRUSSELS 1996-08-22</td>\n",
+ " <td>0</td>\n",
+ " <td>1</td>\n",
+ " <td>Brussels</td>\n",
+ " </tr>\n",
+ " </tbody>\n",
+ "</table>\n",
+ "</div>"
+ ],
+ "text/plain": [
+ " sentence_id sentence beg end \\\n",
+ "0 0000-000 EU rejects German call to boycott British lamb . 0 1 \n",
+ "1 0000-000 EU rejects German call to boycott British lamb . 2 3 \n",
+ "2 0000-000 EU rejects German call to boycott British lamb . 6 7 \n",
+ "3 0000-001 Peter Blackburn 0 2 \n",
+ "4 0000-002 BRUSSELS 1996-08-22 0 1 \n",
+ "\n",
+ " label \n",
+ "0 --NME-- \n",
+ "1 Germany \n",
+ "2 United_Kingdom \n",
+ "3 --NME-- \n",
+ "4 Brussels "
+ ]
+ },
+ "execution_count": 21,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
"df_train.head()"
]
@@ -168,7 +262,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 22,
"metadata": {
"deletable": false,
"nbgrader": {
@@ -184,8 +278,9 @@
},
"outputs": [],
"source": [
+ "\n",
"def evaluation_scores(gold, pred):\n",
- " \"\"\"Print precision, recall, and F1 score.\n",
+ " \"\"\"Return precision, recall, and F1 score.\n",
" \n",
" Arguments:\n",
" gold: The set with the gold-standard values.\n",
@@ -195,8 +290,15 @@
" A tuple or list containing the precision, recall, and F1 values\n",
" (in that order), computed based on the specified sets.\n",
" \"\"\"\n",
- " # YOUR CODE HERE\n",
- " raise NotImplementedError()"
+ " num_correct_pred = len([element for element in pred if element in gold])\n",
+ " num_wrong_pred = len([element for element in pred if element not in gold])\n",
+ " \n",
+ " precision = num_correct_pred / (num_correct_pred + num_wrong_pred)\n",
+ " recall = num_correct_pred / len(gold)\n",
+ " f1 = 2 * (precision * recall) / (precision + recall)\n",
+ "\n",
+ " return precision, recall, f1\n",
+ " "
]
},
{
@@ -210,7 +312,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 23,
"metadata": {
"deletable": false,
"editable": false,
@@ -248,7 +350,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 24,
"metadata": {
"deletable": false,
"editable": false,
@@ -264,7 +366,27 @@
"task": false
}
},
- "outputs": [],
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Precision: 0.500, Recall: 0.333, F1: 0.400\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "<div class=\"alert alert-success\"><strong>Checks have passed!</strong></div>"
+ ],
+ "text/plain": [
+ "<IPython.core.display.HTML object>"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
"source": [
"# Some example NER spans, to illustrate how the evaluation function will be used later\n",
"example_gold = {(\"0000-000\", 0, 1), (\"0000-000\", 2, 3), (\"0000-000\", 6, 7)}\n",
@@ -297,7 +419,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 25,
"metadata": {
"deletable": false,
"editable": false,
@@ -342,7 +464,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 37,
"metadata": {
"deletable": false,
"nbgrader": {
@@ -370,13 +492,13 @@
" position of each span.\n",
" \"\"\"\n",
" # Hint: The Pandas method .itertuples() is useful for iterating over rows in a DataFrame\n",
- " # YOUR CODE HERE\n",
- " raise NotImplementedError()"
+ " for row in df.itertuples():\n",
+ " yield row[1], row[3], row[4]"
]
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 27,
"metadata": {
"deletable": false,
"nbgrader": {
@@ -395,6 +517,8 @@
},
"outputs": [],
"source": [
+ "from spacy.pipeline.ner import DEFAULT_NER_MODEL\n",
+ "\n",
"def pred_spans(df):\n",
" \"\"\"Run and evaluate spaCy's NER.\n",
"\n",
@@ -406,8 +530,13 @@
" triples consisting of the sentence id, start position, and end\n",
" position of each span.\n",
" \"\"\"\n",
- " # YOUR CODE HERE\n",
- " raise NotImplementedError()"
+ " config = {\n",
+ " \"moves\": None,\n",
+ " \"update_with_oracle_cut_size\": 100,\n",
+ " \"model\": DEFAULT_NER_MODEL,\n",
+ " \"incorrect_spans_key\": \"incorrect_spans\",\n",
+ " }\n",
+ " nlp.add_pipe(\"ner\", config=config)\n"
]
},
{
@@ -421,9 +550,21 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 38,
"metadata": {},
- "outputs": [],
+ "outputs": [
+ {
+ "ename": "NameError",
+ "evalue": "name 'spans_dev_pred' is not defined",
+ "output_type": "error",
+ "traceback": [
+ "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+ "\u001b[0;31mNameError\u001b[0m Traceback (most recent call last)",
+ "Cell \u001b[0;32mIn[38], line 6\u001b[0m\n\u001b[1;32m 3\u001b[0m spans_dev_gold \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mset\u001b[39m(gold_spans(df_dev))\n\u001b[1;32m 5\u001b[0m \u001b[38;5;66;03m# Compute and print the evaluation scores\u001b[39;00m\n\u001b[0;32m----> 6\u001b[0m scores \u001b[38;5;241m=\u001b[39m evaluation_scores(spans_dev_gold, \u001b[43mspans_dev_pred\u001b[49m)\n\u001b[1;32m 7\u001b[0m print_evaluation_scores(scores)\n",
+ "\u001b[0;31mNameError\u001b[0m: name 'spans_dev_pred' is not defined"
+ ]
+ }
+ ],
"source": [
"# Produce the spans\n",
"spans_dev_pred = set(pred_spans(df_dev))\n",
@@ -1388,7 +1529,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
- "version": "3.12.7"
+ "version": "3.11.0rc1"
}
},
"nbformat": 4,