picdexer is a picture indexing solution :
- it extracts metadata from pictures
- metadata are indexed on
elasticsearch - pictures can be downsampled and stored using
file-server - it provides a view over indexed data using
kibana
picdexer orchestrates the whole process :
- it configures
elasticsearch - it configures
kibana - it extracts and index pictures metadata
- it can store pictures (possibly resized)
picdexer is released :
- As a compiled binary on the Github project release section. If you have a golang environment, you can compile the projet :
go build -o file-server picdexer.go - As a docker imaged on Docker Hub. If you have a [docker environment], you can build the image :
docker build .
In the picdexer docker image :
- the configuration file is located in
/etc/picdexer/picdexer.json - the "input" folder is located in
/data/picdexer/in/ - the "output" folder (for simulations) is located in
/data/picdexer/out/ - the default "dropzone" folder is
/data/picdexer/dropzone/
The configuration file is a JSON file, here is a complexe example :
{
"loggingLevel": "warn",
"elasticsearch": {
"url": "http://192.168.1.102:9200",
"extractionThreadCount": 4,
"toExtractChannelSize": 50
},
"binary": {
"url": "http://192.168.1.100:8080",
"height":480,
"width":640,
"resizingThreadCount": 4,
"toResizeChannelSize": 4
},
"kibana": {
"url": "http://192.168.1.102:5601"
},
"dropzone": {
"root": "/tmp/foo",
"period": "5s",
"fileChannelSize": 10
}
}loggingLevel(optional) defines the logging level. Values :debug,info(default),warn,error, ...elasticsearch(required if used) configures theelasticsearchconnexion and the metadata extraction processurl(required if documents are pushed) defines theelasticsearchendpointextractionThreadCount(optional, default :4) defines how many thread have to be used to extract medatadatoExtractChannelSize(optimal, default :50) defines the size of the file buffer (go channel) that will be consumed by the extraction goroutines
binary(required if used) configures the interactions withfile-serverto store picturesurl(required if pictures are pushed) defines thefile-serverendpointheightandwidthdefines the target dimension of the pictures that will be stored. If one of the dimension is0then pictures will not be resized (default behaviour).resizingThreadCount(optional, default :4) defines how many thread have to be used to resize picturestoResizeChannelSize(optional, default :4) defines the size of the file buffer (go channel) that will be consumed by the resizing goroutines
kibana(required if user) configures the interaction withkibana(for configuration purpose)url(required if kibana has to be configured) defines thekibanaendpoint
dropzone(required if used) configures dropzoneroot(required) defines the watched folderperioddefines where waiting period between to watching iteration (syntax, ex : 1m, 1h, 30s, ...)fileChannelSize(optional, default :10) defines the size of the buffer (go channel) that will be consumed by indexation
picdexer have several commands that can be used. Each command is dedicated to specific purpose.
If the return code is 0, the command is successfully execute. If any problem happens, the return code will be > 0.
picdexer can setup :
elasticsearchmapping (mapping.json)kibanaindex-pattern, visualizations, dashboards (kibana.ndjson)
0 in kibana. Since there is no official global setting kibana REST API, it has to be setup manually. In the kibana interface, go to Management > Advanced settings > General > Maximum table cell height and set the value to 0.
The elasticsearch and kibana part of the configuration file has to be filled.
- Command line version :
./pcidexer setup -c [configurationFile] - Docker version :
docker run --rm
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./setup.shThe full process command extracts metadata, resize (eventually) and store pictures.
- Command line version :
./picdexer full -c [configurationFile] -d [sourceFolder] - Docker version :
docker run --rm
-v [hostSourceFolder]:/data/picdexer/in
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./full.shThe metadata command deals with elasticsearch metadata indexation.
This command simulates the indexation (dump the elasticsearch documents on stdout).
- Command line version :
./picdexer metadata simulate -c [configurationFile] -d [sourceFolder] - Docker version :
docker run --rm
-v [hostSourceFolder]:/data/picdexer/in
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./metadata_simulate.shThis command indexes metadata to elasticsearch.
- Command line version :
./picdexer metadata index -c [configurationFile] -d [sourceFolder] - Docker version :
docker run --rm
-v [hostSourceFolder]:/data/picdexer/in
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./metadata_index.shThis command extracts all the metadata available for a picture : ./picdexer metadata display -f [pictureFile]
The binary command deals with the picture storage in file-server.
This command simulates the resizing and stores resized file in the filesystem.
- Command line version :
./picdexer binary simulate -c [configurationFile] -d [sourceFolder] -o [outputFolder]. - Docker version :
docker run --rm
-v [hostSourceFolder]:/data/picdexer/in
-v [hostTargetFolder]:/data/picdexer/out
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./binary_simulate.shThis command resizes pictures and stores the resized version in file-server.
- Command line version :
./picdexer binary push -c [configurationFile] -d [sourceFolder]. - Docker version :
docker run --rm
-v [hostSourceFolder]:/data/picdexer/in
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./binary_push.shThis command watches a folder, index, stores pictures and delete files.
- Command line version :
./picdexer dropzone -c [configurationFile] - Docker version :
docker run --rm
-v [hostSourceFolder]:/data/picdexer/in
-v [hostConfigurationFile]:/etc/picdexer/picdexer.json
barasher/picdexer:1.0.0 ./dropzone.sh