From 18719b2b708e3afc921a34aff50d49f8b80157ac Mon Sep 17 00:00:00 2001
From: Sameena-Thabassum <sameena.thaba@freshworks.com>
Date: Tue, 22 Apr 2025 22:36:44 +0530
Subject: [PATCH] Add support for 'basic' chunking strategy to match
 documentation

---
 prepline_general/api/models/form_params.py |  6 ++---
 test_general/api/test_app.py               | 27 ++++++++++++++++++++++
 2 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/prepline_general/api/models/form_params.py b/prepline_general/api/models/form_params.py
index e0f4122e..7d7e00f9 100644
--- a/prepline_general/api/models/form_params.py
+++ b/prepline_general/api/models/form_params.py
@@ -178,11 +178,11 @@ def as_form(
         ] = False,
         # -- chunking options --
         chunking_strategy: Annotated[
-            Optional[Literal["by_title"]],
+            Optional[Literal["by_title", "basic"]],
             Form(
                 title="Chunking Strategy",
-                description="Use one of the supported strategies to chunk the returned elements. Currently supports: by_title",
-                examples=["by_title"],
+                description="Use one of the supported strategies to chunk the returned elements. Currently supports: by_title and basic. Default: None",
+                examples=["by_title", "basic"],
             ),
         ] = None,
         combine_under_n_chars: Annotated[
diff --git a/test_general/api/test_app.py b/test_general/api/test_app.py
index afb743ac..36ca9367 100644
--- a/test_general/api/test_app.py
+++ b/test_general/api/test_app.py
@@ -1153,3 +1153,30 @@ def test_include_slide_notes(monkeypatch, test_default, include_slide_notes, tes
         assert "Here are important notes" == df["text"][0]
     else:
         assert "Here are important notes" != df["text"][0]
+
+def test_basic_chunking_strategy():
+    """
+    Verify that basic chunking strategy works as expected
+    """
+    client = TestClient(app)
+    test_file = Path("sample-docs") / "layout-parser-paper-fast.pdf"
+    response = client.post(
+        MAIN_API_ROUTE,
+        files=[("files", (str(test_file), open(test_file, "rb")))],
+        data={"strategy": "hi_res"},
+    )
+    assert response.status_code == 200
+    response_without_chunking = response.json()
+
+    # chunking
+    response = client.post(
+        MAIN_API_ROUTE,
+        files=[("files", (str(test_file), open(test_file, "rb")))],
+        data={"chunking_strategy": "basic"},
+    )
+    assert response.status_code == 200
+
+    response_with_chunking = response.json()
+    assert len(response_with_chunking) != len(response_without_chunking)
+    assert "CompositeElement" in [element.get("type") for element in response_with_chunking]
+