kevinveenbirkenbach/bill-manager
1
0
Fork 0
mirror of https://github.com/kevinveenbirkenbach/bill-manager.git synced 2024-11-22 01:11:04 +01:00
Code Issues Packages Projects Releases Wiki Activity

Implemented logic for files which don't need ocr

Browse Source
This commit is contained in:
Kevin Veen-Birkenbach 2020-12-10 23:57:29 +01:00
parent c5aa6142ee
commit f58c83fc3f
1 changed files with 13 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
generate.sh
View File

@ -24,7 +24,7 @@ OUTPUT_FOLDER="$1/generated/" || exit 1
if [ "$MODE" = "update" ]; then if [ "$MODE" = "update" ]; then
echo "Updating bills..." echo "Updating bills..."
else else
if [ "$(ls -A "$TMP_FOLDER")" ] if [ "$(ls -A "$OUTPUT_FOLDER")" ]
then then
echo "Cleaning up $OUTPUT_FOLDER..." echo "Cleaning up $OUTPUT_FOLDER..."
rm -v "$OUTPUT_FOLDER"* || exit 1; rm -v "$OUTPUT_FOLDER"* || exit 1;
@ -36,8 +36,17 @@ for origin_file in "$ORIGIN_FOLDER"*.*; do
if [ "$MODE" = "update" ] && [ ! -f "$OUTPUT_FOLDER$(basename "$origin_file")"* ] || [ "$MODE" = "initialize" ]; then if [ "$MODE" = "update" ] && [ ! -f "$OUTPUT_FOLDER$(basename "$origin_file")"* ] || [ "$MODE" = "initialize" ]; then
if [ "$(head -c 4 "$origin_file")" = "%PDF" ]; then if [ "$(head -c 4 "$origin_file")" = "%PDF" ]; then
tmp_file="$TMP_FOLDER$(basename "$origin_file")" tmp_file="$TMP_FOLDER$(basename "$origin_file")"
echo "Generating $tmp_file..." txt_output_file="$OUTPUT_FOLDER$(basename "$origin_file").txt"
pdfimages "$origin_file" "$tmp_file" pdftotext "$origin_file" "$txt_output_file"
content="$(cat "$txt_output_file")"
if [ ${#content} -gt "9" ]
then
echo "Text successfully extracted to $txt_output_file:"
cat "$txt_output_file"
else
echo "Extract images..."
pdfimages "$origin_file" "$tmp_file"
fi
else else
cp -v "$origin_file" "$TMP_FOLDER" cp -v "$origin_file" "$TMP_FOLDER"
fi fi
@ -57,3 +66,4 @@ if [ "$(ls -A "$TMP_FOLDER")" ]
else else
echo "Skipped text generation because $TMP_FOLDER is empty..." echo "Skipped text generation because $TMP_FOLDER is empty..."
fi fi
echo "Cleanup..." && rm -v "$TMP_FOLDER"* && rmdir -v "$TMP_FOLDER";