Files
query-orchestration/internal/document/text/extract_test.go
T
Michael McGuinness fee71e7740 Merged in feature/postprocessing (pull request #114)
Feature/postprocessing

* tests

* passtest

* fixshorttests

* mosttests

* improvingbasedockerfile

* testspeeds

* testing

* host

* canparallel

* clean

* passfullsuite

* singlepagemax

* test

* findfeatures

* findstables

* tbls

* tablestoo

* tablestoo

* lateraltests

* tableloc

* cleanup

* inlinetable

* childids

* cleanup

* tests
2025-04-22 14:40:16 +00:00

182 lines
5.0 KiB
Go

package documenttext
import (
"context"
"fmt"
"io"
"strings"
"testing"
"time"
"queryorchestration/internal/database/repository"
"queryorchestration/internal/serviceconfig"
"queryorchestration/internal/serviceconfig/aws"
"queryorchestration/internal/serviceconfig/objectstore"
"queryorchestration/internal/serviceconfig/queue/querysync"
"queryorchestration/internal/serviceconfig/textract"
objectstoremock "queryorchestration/mocks/objectstore"
queuemock "queryorchestration/mocks/queue"
textractmock "queryorchestration/mocks/textract"
"github.com/aws/aws-sdk-go-v2/service/s3"
"github.com/aws/aws-sdk-go-v2/service/sqs"
awstextract "github.com/aws/aws-sdk-go-v2/service/textract"
"github.com/google/uuid"
"github.com/pashagolub/pgxmock/v3"
"github.com/stretchr/testify/mock"
"github.com/stretchr/testify/require"
)
type DocTextConfig struct {
aws.AWSConfig
serviceconfig.BaseConfig
querysync.QuerySyncConfig
textract.TextractConfig
objectstore.ObjectStoreConfig
}
func TestCreate(t *testing.T) {
ctx := context.Background()
pool, err := pgxmock.NewPool()
require.NoError(t, err)
cfg := &DocTextConfig{}
cfg.DBPool = pool
cfg.DBQueries = repository.New(pool)
mockStore := objectstoremock.NewMockS3Client(t)
cfg.StoreClient = mockStore
mockTextract := textractmock.NewMockTextractClient(t)
cfg.TextractClient = mockTextract
mockQueue := queuemock.NewMockSQSClient(t)
cfg.QueueClient = mockQueue
cfg.QuerySyncURL = "ex"
svc := Service{
cfg: cfg,
}
docId := uuid.New()
t.Run("is not extracted", func(t *testing.T) {
cleanId := uuid.New()
bucket := "bucket"
hash := `"a06d5e0e795adeb6d0b3732330dbcc15"`
textId := uuid.New()
key := objectstore.BucketKey{
ClientID: "aa",
Location: objectstore.Import,
CreatedAt: time.Now().UTC(),
EntityID: cleanId,
}
keyStr := key.String()
textractFile := "../../../assets/sampleGeneration/generated/helloWorld.gen"
textractBaseOut := getTextractFileResponse(t, textractFile)
mimetype := repository.NullCleanmimetype{
Valid: true,
Cleanmimetype: repository.CleanmimetypeApplicationPdf,
}
pool.ExpectQuery("name: IsDocumentTextExtracted :one").WithArgs(docId).WillReturnRows(
pgxmock.NewRows([]string{"isextracted"}).
AddRow(false),
)
pool.ExpectQuery("name: GetCleanEntryByDocId :one").WithArgs(docId).
WillReturnRows(
pgxmock.NewRows([]string{"id", "documentId", "bucket", "key", "version", "hash", "mimetype", "fail", "clientId"}).
AddRow(cleanId, docId, &bucket, &keyStr, int64(123), &hash, mimetype, nil, key.ClientID),
)
bodyMsg := io.NopCloser(strings.NewReader(pdfHelloWorld))
mockStore.EXPECT().
GetObject(
mock.Anything,
mock.MatchedBy(func(in *s3.GetObjectInput) bool {
return *in.Bucket == bucket && *in.Key == keyStr
}),
mock.Anything,
).
Return(&s3.GetObjectOutput{
Body: bodyMsg,
}, nil)
sent := 0
mockTextract.EXPECT().
AnalyzeDocument(
mock.Anything,
mock.MatchedBy(func(in *awstextract.AnalyzeDocumentInput) bool {
if sent == 0 {
sent++
return true
}
return false
}),
mock.Anything,
).
Return(&textractBaseOut["base"][0], nil).Once()
pool.ExpectBegin()
pool.ExpectQuery("name: GetDocumentTextExtractionByHash :one").WithArgs(cleanId, hash).WillReturnRows(
pgxmock.NewRows([]string{"id"}).
AddRow(textId),
)
pool.ExpectExec("name: AddDocumentTextEntry :exec").WithArgs(textId, pgxmock.AnyArg()).
WillReturnResult(pgxmock.NewResult("", 1))
pool.ExpectCommit()
mockQueue.EXPECT().
SendMessage(
mock.Anything,
mock.MatchedBy(func(in *sqs.SendMessageInput) bool {
return *in.QueueUrl == cfg.QuerySyncURL && *in.MessageBody == fmt.Sprintf("{\"id\":\"%s\"}", docId.String())
}),
mock.Anything,
).
Return(&sqs.SendMessageOutput{}, nil)
err = svc.Extract(ctx, docId)
require.NoError(t, err)
})
t.Run("is extracted", func(t *testing.T) {
pool.ExpectQuery("name: IsDocumentTextExtracted :one").WithArgs(docId).WillReturnRows(
pgxmock.NewRows([]string{"isextracted"}).
AddRow(true),
)
mockQueue.EXPECT().
SendMessage(
mock.Anything,
mock.MatchedBy(func(in *sqs.SendMessageInput) bool {
return *in.QueueUrl == cfg.QuerySyncURL && *in.MessageBody == fmt.Sprintf("{\"id\":\"%s\"}", docId.String())
}),
mock.Anything,
).
Return(&sqs.SendMessageOutput{}, nil)
err = svc.Extract(ctx, docId)
require.NoError(t, err)
})
}
func TestInformClean(t *testing.T) {
ctx := context.Background()
mockSQS := queuemock.NewMockSQSClient(t)
cfg := &DocTextConfig{}
cfg.QueueClient = mockSQS
cfg.QuerySyncURL = "/i/am/here"
svc := Service{
cfg: cfg,
}
id := uuid.New()
mockSQS.EXPECT().
SendMessage(
mock.Anything,
mock.MatchedBy(func(in *sqs.SendMessageInput) bool {
return *in.QueueUrl == cfg.QuerySyncURL && *in.MessageBody == fmt.Sprintf("{\"id\":\"%s\"}", id)
}),
mock.Anything,
).
Return(&sqs.SendMessageOutput{}, nil)
err := svc.informExtraction(ctx, id)
require.NoError(t, err)
}