Merge pull request #623 from JohnSnowLabs/fix/hf-datasets-summarization

Fix/hf datasets summarization default prompt

.github/workflows/build_and_test.yml

@@ -25,19 +25,22 @@ jobs:
         with:
           python-version: ${{ matrix.python-version }}
 
-      - name: Cache poetry install
-        uses: actions/cache@v2
-        with:
-          path: ~/.local
-          key: poetry-1.3.1
-
       - uses: snok/install-poetry@v1
         with:
           version: 1.3.1
           virtualenvs-create: true
           virtualenvs-in-project: true
+          installer-parallel: true
+
+      - name: Load cached venv
+        id: cached-poetry-dependencies
+        uses: actions/cache@v3
+        with:
+          path: .venv
+          key: venv-${{ runner.os }}-${{ steps.setup-python.outputs.python-version }}-${{ hashFiles('**/poetry.lock') }}
 
       - name: Install dependencies
+        if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
         run: |
           poetry install --with dev --all-extras
 

langtest/utils/custom_types/helpers.py

@@ -10,7 +10,7 @@
     "quac": "You are an intelligent bot specialized in question answering. Your goal is to provide accurate and concise answers to all the questions without stopping in between. Read the following context and answer each question based on the given information.\n\nContext: {context}\n\nQuestions:\n{question}",
     "narrativeqa": "Context: {context} \nQuestion: {question}\n I've provided a question and context. Answer the given closed-book question based on the provided context. Only answer with words in the context. Answer:",
     "hellaswag": "You are an AI agent that completes sentences and cannot do anything else. You do not repeat the sentence and only continue for one sentence. Complete the following sentence: \n{context}{question}",
-    "sum_prompt": "You are an intelligent Context summarizer. Please read the following context carefully. After understanding its content, create a concise summary, capturing the essential themes and key details. Please ensure that the summary does not end abruptly and remains within the max_tokens word limit. Context: {context}\n\n Summary: ",
+    "default_summarization_prompt": "You are an intelligent Context summarizer. Please read the following context carefully. After understanding its content, create a concise summary, capturing the essential themes and key details. Please ensure that the summary does not end abruptly and remains within the max_tokens word limit. Context: {context}\n\n Summary: ",
     "bbq": "Read the following context carefully and provide a concise answer based solely on the information given. Strictly, do not introduce any new information or make any assumptions. \n\nContext: {context}\nQuestion: {question}\n",
 }
 

langtest/utils/custom_types/sample.py

@@ -677,7 +677,12 @@ def run(self, model, **kwargs):
         Returns:
             bool: True if the sample ran successfully, False otherwise.
         """
-        dataset_name = self.dataset_name.split("-")[0].lower()
+        dataset_name = (
+            self.dataset_name.split("-")[0].lower()
+            if self.dataset_name
+            else "default_summarization_prompt"
+        )
+
         prompt_template = kwargs.get(
             "user_prompt", default_user_prompt.get(dataset_name, "")
         )