mirror/tesseract
1
0
Fork 0
mirror of https://github.com/tesseract-ocr/tesseract.git synced 2024-12-04 01:39:16 +08:00
Code Issues Actions 12 Packages Projects Releases Wiki Activity

Remove --bin_dir option from tesstrain.sh (should use $PATH instead)

Browse Source 
The --bin_dir option to tesstrain.sh is not useful, as $PATH does the
same job much better, so switch to relying on that instead.

This also makes the code a bit more readable, as it removes the need
to refer to binaries as COMMAND_NAME_EXE rather than just command_name.
This commit is contained in:
Nick White 2015-08-26 18:49:14 +01:00 committed by Zdenko Podobný
parent b581c33789
commit 45590ba1c1
2 changed files with 20 additions and 45 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
training/tesstrain.sh
View File

@ -17,7 +17,6 @@
# USAGE: # USAGE:
# #
# tesstrain.sh # tesstrain.sh
# --bin_dir PATH # Location of training program.
# --fontlist FONTS_STR # A plus-separated list of fontnames to train on. # --fontlist FONTS_STR # A plus-separated list of fontnames to train on.
# --fonts_dir FONTS_PATH # Path to font files. # --fonts_dir FONTS_PATH # Path to font files.
# --lang LANG_CODE # ISO 639 code. # --lang LANG_CODE # ISO 639 code.

64
training/tesstrain_utils.sh
View File

@ -41,11 +41,11 @@ err_exit() {
# if the program file is not found. # if the program file is not found.
# Usage: run_command CMD ARG1 ARG2... # Usage: run_command CMD ARG1 ARG2...
run_command() { run_command() {
local cmd=$1 local cmd=`which $1`
shift if [[ -z ${cmd} ]]; then
if [[ ! -x ${cmd} ]]; then err_exit "$1 not found"
err_exit "File ${cmd} not found"
fi fi
shift
tlog "[$(date)] ${cmd} $@" tlog "[$(date)] ${cmd} $@"
${cmd} "$@" 2>&1 1>&2 | tee -a ${LOG_FILE} ${cmd} "$@" 2>&1 1>&2 | tee -a ${LOG_FILE}
# check completion status # check completion status
@ -65,22 +65,6 @@ check_file_readable() {
done done
} }
# Set global path variables that are based on parsed flags.
set_prog_paths() {
if [[ -z ${BINDIR} ]]; then
err_exit "Need to specify location of program files"
fi
CN_TRAINING_EXE=${BINDIR}/cntraining
COMBINE_TESSDATA_EXE=${BINDIR}/combine_tessdata
MF_TRAINING_EXE=${BINDIR}/mftraining
SET_UNICHARSET_PROPERTIES_EXE=${BINDIR}/set_unicharset_properties
SHAPE_TRAINING_EXE=${BINDIR}/shapeclustering
TESSERACT_EXE=${BINDIR}/tesseract
TEXT2IMAGE_EXE=${BINDIR}/text2image
UNICHARSET_EXTRACTOR_EXE=${BINDIR}/unicharset_extractor
WORDLIST2DAWG_EXE=${BINDIR}/wordlist2dawg
}
# Sets the named variable to given value. Aborts if the value is missing or # Sets the named variable to given value. Aborts if the value is missing or
# if it looks like a flag. # if it looks like a flag.
# Usage: parse_value VAR_NAME VALUE # Usage: parse_value VAR_NAME VALUE
@ -105,9 +89,6 @@ parse_flags() {
case ${ARGV[$i]} in case ${ARGV[$i]} in
--) --)
break;; break;;
--bin_dir)
parse_value "BINDIR" ${ARGV[$j]}
i=$j ;;
--fontlist) # Expect a plus-separated list of names --fontlist) # Expect a plus-separated list of names
if [[ -z ${ARGV[$j]} ]] || [[ ${ARGV[$j]:0:2} == "--" ]]; then if [[ -z ${ARGV[$j]} ]] || [[ ${ARGV[$j]:0:2} == "--" ]]; then
err_exit "Invalid value passed to --fontlist" err_exit "Invalid value passed to --fontlist"
@ -152,9 +133,6 @@ parse_flags() {
if [[ -z ${LANG_CODE} ]]; then if [[ -z ${LANG_CODE} ]]; then
err_exit "Need to specify a language --lang" err_exit "Need to specify a language --lang"
fi fi
if [[ -z ${BINDIR} ]]; then
err_exit "Need to specify path to built binaries --bin_dir"
fi
if [[ -z ${LANGDATA_ROOT} ]]; then if [[ -z ${LANGDATA_ROOT} ]]; then
err_exit "Need to specify path to language files --langdata_dir" err_exit "Need to specify path to language files --langdata_dir"
fi fi
@ -167,8 +145,6 @@ parse_flags() {
fi fi
fi fi
set_prog_paths
# Location where intermediate files will be created. # Location where intermediate files will be created.
TRAINING_DIR=${WORKSPACE_DIR}/${LANG_CODE} TRAINING_DIR=${WORKSPACE_DIR}/${LANG_CODE}
# Location of log file for the whole run. # Location of log file for the whole run.
@ -196,7 +172,7 @@ initialize_fontconfig() {
export FONT_CONFIG_CACHE=$(mktemp -d --tmpdir font_tmp.XXXXXXXXXX) export FONT_CONFIG_CACHE=$(mktemp -d --tmpdir font_tmp.XXXXXXXXXX)
local sample_path=${FONT_CONFIG_CACHE}/sample_text.txt local sample_path=${FONT_CONFIG_CACHE}/sample_text.txt
echo "Text" >${sample_path} echo "Text" >${sample_path}
run_command ${TEXT2IMAGE_EXE} --fonts_dir=${FONTS_DIR} \ run_command text2image --fonts_dir=${FONTS_DIR} \
--font="${FONTS[0]}" --outputbase=${sample_path} --text=${sample_path} \ --font="${FONTS[0]}" --outputbase=${sample_path} --text=${sample_path} \
--fontconfig_tmpdir=${FONT_CONFIG_CACHE} --fontconfig_tmpdir=${FONT_CONFIG_CACHE}
} }
@ -224,14 +200,14 @@ generate_font_image() {
fi fi
done done
run_command ${TEXT2IMAGE_EXE} ${common_args} --font="${font}" \ run_command text2image ${common_args} --font="${font}" \
--text=${TRAINING_TEXT} ${TEXT2IMAGE_EXTRA_ARGS} --text=${TRAINING_TEXT} ${TEXT2IMAGE_EXTRA_ARGS}
check_file_readable ${outbase}.box ${outbase}.tif check_file_readable ${outbase}.box ${outbase}.tif
if (( ${EXTRACT_FONT_PROPERTIES} )) && if (( ${EXTRACT_FONT_PROPERTIES} )) &&
[[ -r ${TRAIN_NGRAMS_FILE} ]]; then [[ -r ${TRAIN_NGRAMS_FILE} ]]; then
tlog "Extracting font properties of ${font}" tlog "Extracting font properties of ${font}"
run_command ${TEXT2IMAGE_EXE} ${common_args} --font="${font}" \ run_command text2image ${common_args} --font="${font}" \
--ligatures=false --text=${TRAIN_NGRAMS_FILE} \ --ligatures=false --text=${TRAIN_NGRAMS_FILE} \
--only_extract_font_properties --ptsize=32 --only_extract_font_properties --ptsize=32
check_file_readable ${outbase}.fontinfo check_file_readable ${outbase}.fontinfo
@ -287,7 +263,7 @@ phase_UP_generate_unicharset() {
tlog "\n=== Phase UP: Generating unicharset and unichar properties files ===" tlog "\n=== Phase UP: Generating unicharset and unichar properties files ==="
local box_files=$(ls ${TRAINING_DIR}/*.box) local box_files=$(ls ${TRAINING_DIR}/*.box)
run_command ${UNICHARSET_EXTRACTOR_EXE} -D "${TRAINING_DIR}/" ${box_files} run_command unicharset_extractor -D "${TRAINING_DIR}/" ${box_files}
local outfile=${TRAINING_DIR}/unicharset local outfile=${TRAINING_DIR}/unicharset
UNICHARSET_FILE="${TRAINING_DIR}/${LANG_CODE}.unicharset" UNICHARSET_FILE="${TRAINING_DIR}/${LANG_CODE}.unicharset"
check_file_readable ${outfile} check_file_readable ${outfile}
@ -295,7 +271,7 @@ phase_UP_generate_unicharset() {
XHEIGHTS_FILE="${TRAINING_DIR}/${LANG_CODE}.xheights" XHEIGHTS_FILE="${TRAINING_DIR}/${LANG_CODE}.xheights"
check_file_readable ${UNICHARSET_FILE} check_file_readable ${UNICHARSET_FILE}
run_command ${SET_UNICHARSET_PROPERTIES_EXE} \ run_command set_unicharset_properties \
-U ${UNICHARSET_FILE} -O ${UNICHARSET_FILE} -X ${XHEIGHTS_FILE} \ -U ${UNICHARSET_FILE} -O ${UNICHARSET_FILE} -X ${XHEIGHTS_FILE} \
--script_dir=${LANGDATA_ROOT} --script_dir=${LANGDATA_ROOT}
check_file_readable ${XHEIGHTS_FILE} check_file_readable ${XHEIGHTS_FILE}
@ -323,7 +299,7 @@ phase_D_generate_dawg() {
if [[ -s ${WORDLIST_FILE} ]]; then if [[ -s ${WORDLIST_FILE} ]]; then
tlog "Generating word Dawg" tlog "Generating word Dawg"
check_file_readable ${UNICHARSET_FILE} check_file_readable ${UNICHARSET_FILE}
run_command ${WORDLIST2DAWG_EXE} -r 1 ${WORDLIST_FILE} ${WORD_DAWG} \ run_command wordlist2dawg -r 1 ${WORDLIST_FILE} ${WORD_DAWG} \
${UNICHARSET_FILE} ${UNICHARSET_FILE}
check_file_readable ${WORD_DAWG} check_file_readable ${WORD_DAWG}
@ -335,13 +311,13 @@ phase_D_generate_dawg() {
if [[ -s ${freq_wordlist_file} ]]; then if [[ -s ${freq_wordlist_file} ]]; then
check_file_readable ${UNICHARSET_FILE} check_file_readable ${UNICHARSET_FILE}
tlog "Generating frequent-word Dawg" tlog "Generating frequent-word Dawg"
run_command ${WORDLIST2DAWG_EXE} -r 1 ${freq_wordlist_file} \ run_command wordlist2dawg -r 1 ${freq_wordlist_file} \
${FREQ_DAWG} ${UNICHARSET_FILE} ${FREQ_DAWG} ${UNICHARSET_FILE}
check_file_readable ${FREQ_DAWG} check_file_readable ${FREQ_DAWG}
fi fi
# Punctuation DAWG # Punctuation DAWG
# -r arguments to WORDLIST2DAWG_EXE denote RTL reverse policy # -r arguments to wordlist2dawg denote RTL reverse policy
# (see Trie::RTLReversePolicy enum in third_party/tesseract/dict/trie.h). # (see Trie::RTLReversePolicy enum in third_party/tesseract/dict/trie.h).
# We specify 0/RRP_DO_NO_REVERSE when generating number DAWG, # We specify 0/RRP_DO_NO_REVERSE when generating number DAWG,
# 1/RRP_REVERSE_IF_HAS_RTL for freq and word DAWGS, # 1/RRP_REVERSE_IF_HAS_RTL for freq and word DAWGS,
@ -356,20 +332,20 @@ phase_D_generate_dawg() {
PUNC_FILE="${LANGDATA_ROOT}/common.punc" PUNC_FILE="${LANGDATA_ROOT}/common.punc"
fi fi
check_file_readable ${PUNC_FILE} check_file_readable ${PUNC_FILE}
run_command ${WORDLIST2DAWG_EXE} -r ${punc_reverse_policy} \ run_command wordlist2dawg -r ${punc_reverse_policy} \
${PUNC_FILE} ${PUNC_DAWG} ${UNICHARSET_FILE} ${PUNC_FILE} ${PUNC_DAWG} ${UNICHARSET_FILE}
check_file_readable ${PUNC_DAWG} check_file_readable ${PUNC_DAWG}
# Numbers DAWG # Numbers DAWG
if [[ -s ${NUMBERS_FILE} ]]; then if [[ -s ${NUMBERS_FILE} ]]; then
run_command ${WORDLIST2DAWG_EXE} -r 0 \ run_command wordlist2dawg -r 0 \
${NUMBERS_FILE} ${NUMBER_DAWG} ${UNICHARSET_FILE} ${NUMBERS_FILE} ${NUMBER_DAWG} ${UNICHARSET_FILE}
check_file_readable ${NUMBER_DAWG} check_file_readable ${NUMBER_DAWG}
fi fi
# Bigram dawg # Bigram dawg
if [[ -s ${WORD_BIGRAMS_FILE} ]]; then if [[ -s ${WORD_BIGRAMS_FILE} ]]; then
run_command ${WORDLIST2DAWG_EXE} -r 1 \ run_command wordlist2dawg -r 1 \
${WORD_BIGRAMS_FILE} ${BIGRAM_DAWG} ${UNICHARSET_FILE} ${WORD_BIGRAMS_FILE} ${BIGRAM_DAWG} ${UNICHARSET_FILE}
check_file_readable ${BIGRAM_DAWG} check_file_readable ${BIGRAM_DAWG}
fi fi
@ -401,7 +377,7 @@ phase_E_extract_features() {
tlog "Using TESSDATA_PREFIX=${TESSDATA_PREFIX}" tlog "Using TESSDATA_PREFIX=${TESSDATA_PREFIX}"
local counter=0 local counter=0
for img_file in ${img_files}; do for img_file in ${img_files}; do
run_command ${TESSERACT_EXE} ${img_file} ${img_file%.*} \ run_command tesseract ${img_file} ${img_file%.*} \
${box_config} ${config} & ${box_config} ${config} &
let counter=counter+1 let counter=counter+1
let rem=counter%par_factor let rem=counter%par_factor
@ -423,7 +399,7 @@ phase_C_cluster_prototypes() {
tlog "\n=== Phase C: Clustering feature prototypes (cnTraining) ===" tlog "\n=== Phase C: Clustering feature prototypes (cnTraining) ==="
local out_normproto=$1 local out_normproto=$1
run_command ${CN_TRAINING_EXE} -D "${TRAINING_DIR}/" \ run_command cntraining -D "${TRAINING_DIR}/" \
$(ls ${TRAINING_DIR}/*.tr) $(ls ${TRAINING_DIR}/*.tr)
check_file_readable ${TRAINING_DIR}/normproto check_file_readable ${TRAINING_DIR}/normproto
@ -443,7 +419,7 @@ phase_S_cluster_shapes() {
font_props=${font_props}" -X ${TRAINING_DIR}/${LANG_CODE}.xheights" font_props=${font_props}" -X ${TRAINING_DIR}/${LANG_CODE}.xheights"
fi fi
run_command ${SHAPE_TRAINING_EXE} \ run_command shapeclustering \
-D "${TRAINING_DIR}/" \ -D "${TRAINING_DIR}/" \
-U ${TRAINING_DIR}/${LANG_CODE}.unicharset \ -U ${TRAINING_DIR}/${LANG_CODE}.unicharset \
-O ${TRAINING_DIR}/${LANG_CODE}.mfunicharset \ -O ${TRAINING_DIR}/${LANG_CODE}.mfunicharset \
@ -464,7 +440,7 @@ phase_M_cluster_microfeatures() {
font_props=${font_props}" -X ${TRAINING_DIR}/${LANG_CODE}.xheights" font_props=${font_props}" -X ${TRAINING_DIR}/${LANG_CODE}.xheights"
fi fi
run_command ${MF_TRAINING_EXE} \ run_command mftraining \
-D "${TRAINING_DIR}/" \ -D "${TRAINING_DIR}/" \
-U ${TRAINING_DIR}/${LANG_CODE}.unicharset \ -U ${TRAINING_DIR}/${LANG_CODE}.unicharset \
-O ${TRAINING_DIR}/${LANG_CODE}.mfunicharset \ -O ${TRAINING_DIR}/${LANG_CODE}.mfunicharset \
@ -524,7 +500,7 @@ make__traineddata() {
fi fi
# Compose the traineddata file. # Compose the traineddata file.
run_command ${COMBINE_TESSDATA_EXE} ${TRAINING_DIR}/${LANG_CODE}. run_command combine_tessdata ${TRAINING_DIR}/${LANG_CODE}.
# Copy it to the output dir, overwriting only if allowed by the cmdline flag. # Copy it to the output dir, overwriting only if allowed by the cmdline flag.
if [[ ! -d ${OUTPUT_DIR} ]]; then if [[ ! -d ${OUTPUT_DIR} ]]; then