Files
query-orchestration/api/docTextRunner/runner_test.go
T
jay brown a9d81a1094 remove parallel
remove parallel to stabilize ci/cd
2025-07-22 10:29:44 -07:00

204 lines
4.6 KiB
Go

package doctextrunner_test
import (
"fmt"
"io"
"regexp"
"strings"
"testing"
"time"
doctextrunner "queryorchestration/api/docTextRunner"
"queryorchestration/internal/database/repository"
documenttext "queryorchestration/internal/document/text"
"queryorchestration/internal/server/runner"
"queryorchestration/internal/serviceconfig/aws"
"queryorchestration/internal/serviceconfig/objectstore"
"queryorchestration/internal/serviceconfig/queue/querysync"
"queryorchestration/internal/serviceconfig/textract"
"queryorchestration/internal/test"
objectstoremock "queryorchestration/mocks/objectstore"
textractmock "queryorchestration/mocks/textract"
"github.com/aws/aws-sdk-go-v2/service/s3"
awstextract "github.com/aws/aws-sdk-go-v2/service/textract"
"github.com/google/uuid"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/mock"
"github.com/stretchr/testify/require"
)
type DocTextConfig struct {
runner.BaseConfig[doctextrunner.Body]
aws.AWSConfig
querysync.QuerySyncConfig
textract.TextractConfig
objectstore.ObjectStoreConfig
}
func TestDocTextRunner(t *testing.T) {
cfg := &DocTextConfig{}
test.CreateDB(t, cfg)
acfg := test.CreateAWSContainer(t, cfg)
test.SetQueueClient(t, t.Context(), cfg, acfg.ExternalEndpoint)
cfg.QuerySyncURL = test.CreateQueue(t, cfg, test.QuerySyncRunnerName)
mockStore := objectstoremock.NewMockS3Client(t)
cfg.StoreClient = mockStore
mockTextract := textractmock.NewMockTextractClient(t)
cfg.TextractClient = mockTextract
runner := doctextrunner.New(&doctextrunner.Services{
Text: documenttext.New(cfg),
})
err := cfg.GetDBQueries().CreateClient(t.Context(), &repository.CreateClientParams{
Clientid: "clientid",
Name: "client_name",
})
require.NoError(t, err)
docId, err := cfg.GetDBQueries().CreateDocument(t.Context(), &repository.CreateDocumentParams{
Clientid: "clientid",
Hash: "hash",
})
require.NoError(t, err)
part := uint16(1)
filetype := "pdf"
key := objectstore.BucketKey{
CreatedAt: time.Now().UTC(),
ClientID: "clientid",
EntityID: uuid.New(),
Location: objectstore.Import,
Part: &part,
FileType: &filetype,
}
err = cfg.GetDBQueries().AddDocumentEntry(t.Context(), &repository.AddDocumentEntryParams{
Documentid: docId,
Bucket: "bucket",
Key: key.String(),
})
require.NoError(t, err)
bucket := "bucket"
hash := "hash"
keyStr := key.String()
cleanId, err := cfg.GetDBQueries().AddDocumentClean(t.Context(), &repository.AddDocumentCleanParams{
Documentid: docId,
Bucket: &bucket,
Hash: &hash,
Key: &keyStr,
Mimetype: repository.NullCleanmimetype{
Valid: true,
Cleanmimetype: repository.CleanmimetypeApplicationPdf,
},
})
require.NoError(t, err)
err = cfg.GetDBQueries().AddDocumentCleanEntry(t.Context(), &repository.AddDocumentCleanEntryParams{
Cleanid: cleanId,
Version: 1,
})
require.NoError(t, err)
bodyMsg := io.NopCloser(strings.NewReader(pdfHelloWorld))
mockStore.EXPECT().
GetObject(
mock.Anything,
mock.MatchedBy(func(in *s3.GetObjectInput) bool {
return *in.Bucket == "bucket" && *in.Key == key.String()
}),
mock.Anything,
).
Return(&s3.GetObjectOutput{
Body: bodyMsg,
}, nil)
filename := "../../assets/textract/helloWorld.gen"
textractBaseOut := test.GetTextractFileResponse(t, filename)
mockTextract.EXPECT().
AnalyzeDocument(
mock.Anything,
mock.MatchedBy(func(in *awstextract.AnalyzeDocumentInput) bool {
return true
}),
mock.Anything,
).
Return(textractBaseOut[0]["base"], nil).Once()
mockStore.EXPECT().
PutObject(
mock.Anything,
mock.MatchedBy(func(in *s3.PutObjectInput) bool {
return *in.Bucket == "bucket"
}),
mock.Anything,
).
Return(&s3.PutObjectOutput{}, nil)
doc := doctextrunner.Body{
DocumentID: docId,
}
assert.True(t, runner.Process(t.Context(), doc))
test.AssertMessageBody(t, cfg, cfg.GetQuerySyncURL(), regexp.MustCompile(fmt.Sprintf("{\"id\":\"%s\"}", docId.String())))
}
const pdfHelloWorld = `%PDF-1.4
%
1 0 obj
<<
/Type /Catalog
/Pages 2 0 R
/Version /1.4
>>
endobj
2 0 obj
<<
/Type /Pages
/Kids [3 0 R]
/Count 1
>>
endobj
3 0 obj
<<
/Type /Page
/Parent 2 0 R
/MediaBox [0 0 612 792]
/Resources <<
/Font <<
/F1 <<
/Type /Font
/Subtype /Type1
/BaseFont /Helvetica
>>
>>
>>
/Contents 4 0 R
>>
endobj
4 0 obj
<<
/Length
44
>>
stream
BT
/F1 24 Tf
100 700 Td
(Hello World!) Tj
ET
endstream
endobj
xref
0 5
0000000000 65535 f
0000000015 00000 n
0000000086 00000 n
0000000151 00000 n
0000000376 00000 n
trailer
<<
/Size 5
/Root 1 0 R
>>
startxref
472
%%EOF`