diff --git a/.dockerignore b/.dockerignore index af41ed2..eba6ef7 100644 --- a/.dockerignore +++ b/.dockerignore @@ -1,8 +1,5 @@ *.log *.json *.md -*.pdf -*.xml -*.jpg *.dtd out/* \ No newline at end of file diff --git a/Dockerfile b/Dockerfile index 4dba562..2f46371 100644 --- a/Dockerfile +++ b/Dockerfile @@ -1,9 +1,9 @@ FROM golang:1.17-alpine -RUN apk add poppler-utils +RUN apk add poppler-utils gcc musl-dev # build WORKDIR /app -COPY * /app/ +ADD . /app/ RUN go build . # Create user diff --git a/main.go b/main.go index d38da3a..4b7936b 100644 --- a/main.go +++ b/main.go @@ -48,7 +48,7 @@ xml LogMessageXML } -var queueForConcurrent = make(chan struct{}, 1100) +var queueForConcurrent = make(chan struct{}, 1500) var wg sync.WaitGroup var numberFiles int = 0 var corpusPath = flag.String("p", "", "Corpus path") diff --git a/pdf_test.go b/pdf_test.go index 55f2ef3..3987bf3 100644 --- a/pdf_test.go +++ b/pdf_test.go @@ -1,7 +1,6 @@ package main import ( - "reflect" "testing" "github.com/stretchr/testify/assert" @@ -20,17 +19,10 @@ // Test metadata func TestMetadata(t *testing.T) { result := getMetadata(&pdfData) - metadata := LogMessagePDF{ - pdfPageTotal: 1, - pdfWordByPage: 0, - pdfWordCount: 0, - Author: "manshanden", - Creator: "PScript5.dll Version 5.2", - CreationDate: "Thu Mar 4 13:08:00 2010 CET", - pdfFormatVersion: "1.4", - pdfError: "", - } - assert.Equal(t, reflect.DeepEqual(metadata, result.pdf), true, "Metadata is equal") + assert.Equal(t, result.pdf.Author, "manshanden", "author is equal") + assert.Equal(t, result.pdf.Creator, "PScript5.dll Version 5.2", "Creator is equal") + assert.Equal(t, result.pdf.pdfFormatVersion, "1.4", "pdfFormatVersion is equal") + assert.Equal(t, result.pdf.pdfError, "", "No error") } // test number word