fix(cli): Load question from config.

README.md

@@ -51,7 +51,7 @@ pip install structured-qa
 structured-qa \
 --question "What optimizer was used to train the model?" \
 --input_file "example_data/1706.03762v7.pdf" \
---output_folder "example_outputs/1706.03762v7.pdf"
+--output_dir "example_outputs/1706.03762v7.pdf"
 ```
 
 ### Graphical Interface App

docs/cli.md

@@ -15,7 +15,7 @@ Or provide values to the arguments directly:
 structured-qa \
 --question "What learning rate was used?" \
 --input_file "example_data/1706.03762v7.pdf" \
---output_folder "example_outputs/1706.03762v7.pdf"
+--output_dir "example_outputs/1706.03762v7.pdf"
 ```
 
 ---

docs/customization.md

@@ -6,6 +6,8 @@ This Blueprint is designed to be flexible and easily adaptable to your specific
 
 ## 🖋️ **Customizable Parameters**
 
+-- **`question`**: The question to be answered.
+
 - **`input_file`**: The input file specifies the document to be processed. Supports the `pdf` format.
 
 - **`output_dir`**: Path to the output directory where the extracted sections will be saved.

example_data/config.yaml

@@ -1,3 +1,4 @@
+question: "What optimizer was used to train the model?"
 input_file: example_data/1706.03762v7.pdf
 output_dir: example_outputs/1706.03762v7.pdf
 model: "bartowski/Qwen2.5-3B-Instruct-GGUF/Qwen2.5-3B-Instruct-f16.gguf"

src/structured_qa/cli.py

@@ -12,7 +12,7 @@
 
 @logger.catch(reraise=True)
 def structured_qa(
-    question: str,
+    question: str | None = None,
     input_file: str | None = None,
     output_dir: str | None = None,
     model: str
@@ -57,6 +57,7 @@ def structured_qa(
     else:
         Path(output_dir).mkdir(exist_ok=True, parents=True)
         config = Config(
+            question=question,
             input_file=input_file,
             output_dir=output_dir,
             model=model,

src/structured_qa/config.py

@@ -60,6 +60,7 @@ def answer_prompt(value):
 
 
 class Config(BaseModel):
+    question: str
     input_file: FilePath
     output_dir: DirectoryPath
     model: Annotated[str, AfterValidator(validate_model)]