chore: preload models in Dockerfile, bump VERSION, add pytest and unit tests

Dockerfile

@@ -16,6 +16,11 @@ RUN pip install --upgrade pip && \
 
 COPY . .
 
+# Pre-download commonly used models to avoid long first-request cold starts.
+# These lines increase the image size but significantly reduce latency on first API call.
+RUN python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('paraphrase-multilingual-MiniLM-L12-v2')"
+RUN python -c "from transformers import pipeline; pipeline('sentiment-analysis', model='nlptown/bert-base-multilingual-uncased-sentiment')"
+
 EXPOSE 8000
 
 CMD ["python", "run.py"]

VERSION

@@ -1 +1 @@
-0.1.0
+0.2.0

requirements.txt

@@ -16,3 +16,6 @@ google-generativeai==0.6.0
 pyarrow==14.0.2
 tiktoken==0.7.0
 
+# Dev / test dependencies
+pytest==7.4.0
+

tests/test_analysis.py

@@ -0,0 +1,27 @@
+import pandas as pd
+from app.analysis import detect_query_type, count_keyword_rows, THANKS_KEYWORDS, COMPLAINT_KEYWORDS
+
+
+def test_detect_query_type_thanks():
+    t = "כמה משתמשים כתבו תודה על השירות"
+    qtype, target = detect_query_type(t)
+    assert qtype == "count_thanks"
+
+
+def test_detect_query_type_complaint():
+    t = "כמה אנשים מדווחים על תקלה במערכת"
+    qtype, target = detect_query_type(t)
+    assert qtype == "count_complaint"
+
+
+def test_count_keyword_rows_counts_thanks():
+    df = pd.DataFrame({"Text": ["תודה על השירות", "לא טוב", "תודה רבה!", "שדרוג נחמד"]})
+    cnt = count_keyword_rows(df, THANKS_KEYWORDS, text_column="Text")
+    assert cnt == 2
+
+
+def test_count_keyword_rows_counts_complaints():
+    df = pd.DataFrame({"Text": ["אין שגיאה", "תיקון נדרש", "יש תקלה", "נכשל בעבודה"]})
+    # Notice: keywords are Hebrew complaint variants; expect matches >=1
+    cnt = count_keyword_rows(df, COMPLAINT_KEYWORDS, text_column="Text")
+    assert cnt >= 1