Expanding the e2e test (#25)

* Expanding the e2e test

It now includes training, testing, and evaluation

* Making the training thread count explicit

test_e2e

@@ -17,9 +17,19 @@ OUTPUT_DIR=$(mktemp -d)
 CRF_TRAINING_FILE="${OUTPUT_DIR}/training_data.crf"
 CRF_TESTING_FILE="${OUTPUT_DIR}/testing_data.crf"
 
+CRF_LEARN_TEMPLATE=template_file
+# This needs to be explicit so that there is consistent training between
+# different machines.
+CRF_TRAINING_THREADS=2
+CRF_MODEL_FILE="${OUTPUT_DIR}/crf_model"
+
+TESTING_OUTPUT_FILE="${OUTPUT_DIR}/testing_output"
+EVAL_OUTPUT_FILE="${OUTPUT_DIR}/eval_output"
+
 GOLDEN_DIR=tests/golden
 GOLDEN_CRF_TRAINING_FILE="${GOLDEN_DIR}/training_data.crf"
 GOLDEN_CRF_TESTING_FILE="${GOLDEN_DIR}/testing_data.crf"
+GOLDEN_EVAL_OUTPUT_FILE="${GOLDEN_DIR}/eval_output"
 
 bin/generate_data \
   --data-path="$LABELLED_DATA_FILE" \
@@ -35,4 +45,17 @@ bin/generate_data \
 
 diff --context=2 "$GOLDEN_CRF_TESTING_FILE" "$CRF_TESTING_FILE"
 
+crf_learn \
+  --thread="$CRF_TRAINING_THREADS" \
+  "$CRF_LEARN_TEMPLATE" "$CRF_TESTING_FILE" "$CRF_MODEL_FILE"
+
+crf_test \
+  --model="$CRF_MODEL_FILE" \
+  "$CRF_TESTING_FILE" > "$TESTING_OUTPUT_FILE"
+
+python bin/evaluate.py "$TESTING_OUTPUT_FILE" > "$EVAL_OUTPUT_FILE"
+cat "$EVAL_OUTPUT_FILE"
+
+diff "$GOLDEN_EVAL_OUTPUT_FILE" "$EVAL_OUTPUT_FILE"
+
 rm -rf $OUTPUT_DIR

tests/golden/eval_output

@@ -0,0 +1,10 @@
+
+Sentence-Level Stats:
+	correct:  1858
+	total:  1999
+	% correct:  92.9464732366
+
+Word-Level Stats:
+	correct: 11249
+	total: 11450
+	% correct: 98.2445414847