fix minor bug in natural language description system. also add natura…

…l language description to the tutorial.
trane-dev · Feb 5, 2018 · 0df4e65 · 0df4e65
1 parent dd88524
commit 0df4e65
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 9 deletions.
diff --git a/Tutorial.ipynb b/Tutorial.ipynb
@@ -95,14 +95,28 @@
   {
    "cell_type": "code",
    "execution_count": 4,
-   "metadata": {
-    "collapsed": true
-   },
-   "outputs": [],
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['For each taxi_id, predict the first fare, after trip_id 0.',\n",
+       " 'For each taxi_id, predict the first fare, after trip_id 0.',\n",
+       " 'For each taxi_id, predict the first fare, after trip_id 0.']"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
    "source": [
     "prediction_problems_json = trane.prediction_problems_to_json_file(\n",
     "    probs, table_meta, entity_id_column, label_generating_column, time_column, \"prediction_problems.json\")\n",
     "\n",
+    "trane.generate_nl_description(\n",
+    "    probs, table_meta, entity_id_column, label_generating_column, time_column, trane.ConstantIntegerCutoffTimes(0))\n",
+    "\n",
     "# with open(\"prediction_problems.json\", \"w\") as f:\n",
     "#     json.dump(json.loads(prediction_problems_json), f, indent=4, separators=(',', ': '))"
    ]
@@ -301,7 +315,9 @@
   {
    "cell_type": "code",
    "execution_count": 6,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "entity_to_data_and_cutoff_dict = trane.ConstantIntegerCutoffTimes(0).generate_cutoffs(entity_to_data_dict)"

diff --git a/trane/utils/generate_nl_description.py b/trane/utils/generate_nl_description.py
@@ -36,7 +36,7 @@ def generate_nl_description(
     """
 
     def description(prob):
-        return "For each {col}, predict{dataop_des},{filter_des},{cutoff_des}.".format(
+        return "For each {col}, predict{dataop_des}{filter_des}{cutoff_des}.".format(
             col=entity_id_column,
             dataop_des=dataop_description(prob),
             filter_des=filter_description(prob),
@@ -97,9 +97,9 @@ def aggop_description():
 
     def cutoff_description(prob):
         if isinstance(cutoff_strategy, ConstantIntegerCutoffTimes):
-            return " after {col} {cutoff}".format(col=time_column, cutoff=cutoff_strategy.integer_cutoff)
+            return ", after {col} {cutoff}".format(col=time_column, cutoff=cutoff_strategy.integer_cutoff)
         elif isinstance(cutoff_strategy, ConstantDatetimeCutoffTimes):
-            return " after {col} {cutoff}".format(col=time_column, cutoff=cutoff_strategy.datetime_cutoff)
+            return ", after {col} {cutoff}".format(col=time_column, cutoff=cutoff_strategy.datetime_cutoff)
         else:
             raise " (unknown cutoff time)"
 
@@ -113,7 +113,7 @@ def filter_description(prob):
         filter_op = prob.operations[0]
         if isinstance(filter_op, AllFilterOp):
             return ""
-        return " with {col} {op} {threshold}".format(
+        return ", with {col} {op} {threshold}".format(
             col=filter_op.column_name, 
             op=filter_op_str_dict[type(filter_op)],
             threshold=filter_op.param_values['threshold'])