test: enable kubernetes tests for python-language-pdf2parquet

also updates to latest from data-prep-kit

Signed-off-by: Nick Mitchell <[email protected]>

Author: starpit

pkg/be/local/shell/ok.go

@@ -20,7 +20,9 @@ func isCompatibleImage(image string) bool {
 func IsCompatible(c llir.ShellComponent) error {
 	switch {
 	case c.Application.Spec.Image != "" && !isCompatibleImage(c.Application.Spec.Image):
-		return fmt.Errorf("Unable to target the local backend because a component '%s' needs to run in a container: %s", c.C(), c.Application.Spec.Image)
+		if slices.IndexFunc(c.Application.Spec.Needs, func(need hlir.Needs) bool { return need.Requirements != "" }) < 0 {
+			return fmt.Errorf("Unable to target the local backend because a component '%s' needs to run in a container: %s", c.C(), c.Application.Spec.Image)
+		}
 
 	case c.Application.Spec.SecurityContext.RunAsUser != 0 ||
 		c.Application.Spec.SecurityContext.RunAsGroup != 0 ||

tests/tests/python-language-pdf2parquet/pail/app.yaml

@@ -5,6 +5,9 @@ metadata:
 spec:
   image: docker.io/python:3.12
   command: python3 ./main.py
+  minSize: sm
+  env:
+    USE_NNPACK: '0' # otherwise torch fails with "Could not initialize NNPACK! Reason: Unsupported hardware" on ARM (lack of AVX instructions)
   needs:
     - name: python
       version: latest

tests/tests/python-language-pdf2parquet/pail/data/requirements.txt

@@ -1,7 +1,7 @@
-docling-core==1.3.0
-docling-ibm-models==1.1.7
-deepsearch-glm==0.21.0
-docling==1.11.0
+docling-core==2.3.0
+docling-ibm-models==2.0.3
+deepsearch-glm==0.26.1
+docling==2.3.1
 filetype >=1.2.0, <2.0.0
 
 # we can probably update to 18+, but we will have to re-generate expected output as pyarrow 18 seems to have resulted in a binary format change