refactoring: splitting source data from encoded data in data fixture

test/fixtures/input.py

@@ -10,6 +10,7 @@ from pyinfra.server.normalization import normalize_item
 from pyinfra.server.packing import pack, unpack
 from pyinfra.utils.func import star, lift, lstarlift
 from test.utils.image import image_to_bytes
+from test.utils.pdf import pdf_stream
 
 
 @pytest.fixture
@@ -21,17 +22,43 @@ def data(data_type, pdf):
 
 
 @pytest.fixture
-def input_data_items(item_type, n_items, pdf):
+def input_data_items(unencoded_input_data, input_data_encoder):
+    return input_data_encoder(unencoded_input_data)
+
+
+@pytest.fixture
+def unencoded_input_data(item_type, unencoded_strings, unencoded_images, unencoded_pdfs):
     if item_type == "string":
-        return [bytes(f"content{i}", encoding="utf8") for i in range(n_items)]
+        return unencoded_strings
     elif item_type == "image":
-        return images(n_items)
+        return unencoded_images
     elif item_type == "pdf":
-        return [pdf] * n_items
+        return unencoded_pdfs
     else:
         raise ValueError(f"Unknown item type {item_type}")
 
 
+@pytest.fixture
+def input_data_encoder(item_type):
+    if item_type == "string":
+        return strings_to_bytes
+    elif item_type == "image":
+        return images_to_bytes
+    elif item_type == "pdf":
+        return pdfs_to_bytes
+    else:
+        raise ValueError(f"Unknown item type {item_type}")
+
+
+@pytest.fixture
+def unencoded_pdfs(n_items, unencoded_pdf):
+    return [unencoded_pdf] * n_items
+
+
+def pdfs_to_bytes(unencoded_pdfs):
+    return [pdf_stream(pdf) for pdf in unencoded_pdfs]
+
+
 @pytest.fixture
 def target_data_items(input_data_items, core_operation):
 
@@ -43,6 +70,15 @@ def target_data_items(input_data_items, core_operation):
     return expected
 
 
+@pytest.fixture
+def unencoded_strings(n_items):
+    return [f"content{i}" for i in range(n_items)]
+
+
+def strings_to_bytes(strings):
+    return [bytes(s, encoding="utf8") for s in strings]
+
+
 @pytest.fixture
 def targets(input_data_items, operation, metadata):
 
@@ -87,8 +123,13 @@ def input_batch(n_items):
     return np.random.random_sample(size=(n_items, 3, 30, 30))
 
 
-def images(n_items):
-    return lmap(compose(image_to_bytes, array_to_image), input_batch(n_items))
+@pytest.fixture
+def unencoded_images(n_items):
+    return lmap(array_to_image, input_batch(n_items))
+
+
+def images_to_bytes(images):
+    return lmap(image_to_bytes, images)
 
 
 @pytest.fixture

test/fixtures/pdf.py

@@ -1,13 +1,11 @@
 import fpdf
 import pytest
 
-from test.utils.pdf import pdf_stream
-
 
 @pytest.fixture
-def pdf(n_pages):
+def unencoded_pdf(n_pages):
     pdf = fpdf.FPDF(unit="pt")
     for _ in range(n_pages):
         pdf.add_page()
 
-    return pdf_stream(pdf)
+    return pdf