Here's a patch that leverages the fact that SimplePostTool is already in solr-core-*.jar. With this patch you can do "bin/post /some/directory". This is just a quick proof-of-concept; there are TODO's in the script to tackle to flesh this out fully.

I'm doing this kind of thing to demonstrate Solr with the post.jar tool:

java -classpath example/solr-webapp/webapp/WEB-INF/lib/solr-core-*.jar -Ddata=web -Drecursive=1 -Ddelay=1 -Dc=gettingstarted -Dauto org.apache.solr.util.SimplePostTool $@

That's the kind of thing we can get bin/post to do cleanly for some very common use cases (file, web, data files).

As part of the work I'm doing for SOLR-3619, I'm also invoking the post tool using:

"$JAVA" -Durl=http://localhost:$SOLR_PORT/solr/$EXAMPLE/update -jar $SOLR_TIP/example/exampledocs/post.jar $SOLR_TIP/example/exampledocs/*.xml

Of course, this complexity should be hidden behind the simple: bin/solr post command

Also as part of the work in SOLR-3619, the script will be able to auto-detect the port a local Solr is listening too, so that users don't have to do things like:

bin/solr post -url http://localhost:8983/solr ...

Current patch works like this:

# Usage:
#  bin/post <collection> http://lucidworks.com [depth=1] [delay=1]
#  bin/post <collection> ~/Documents
#  bin/post <collection> files*.csv
#  bin/post <collection> files*.xml
#  bin/post <collection> files*.json

arbitrary parameters, after the second (file or URL) parameter, are automatically made -D system properties when invoking SimplePostTool.

With the latest patch, my example above becomes

bin/post gettingstarted http://lucidworks.com recursive=1 delay=1

And Tim's example becomes

bin/post $EXAMPLE *.xml port=$SOLR_PORT

+1

Looks great!

Error checking is the tough part. Latest patch requires this syntax "bin/post <collection> <path or url> [optional params passed to SPT]", but if the user omits the collection name what then? Ugly error currently. I suppose it could check for the existence of the collection and issue a clean error message.

Any objections or thoughts about committing it basically like this and iterate? I don't plan on making a comparable Windows version of this myself, but patches welcome on that front.

I tried indexing *.xml and it just ended up indexing 1 file as there seems to be a problem with the expansion:

solr $ bin/post gettingstarted example/exampledocs/*.xml
Collection: gettingstarted
PATH
java -classpath example/solr-webapp/webapp/WEB-INF/lib/solr-core-6.0.0-SNAPSHOT.jar -Dc=gettingstarted -Dexample/exampledocs/hd.xml -Dexample/exampledocs/ipod_other.xml -Dexample/exampledocs/ipod_video.xml -Dexample/exampledocs/manufacturers.xml -Dexample/exampledocs/mem.xml -Dexample/exampledocs/money.xml -Dexample/exampledocs/monitor.xml -Dexample/exampledocs/monitor2.xml -Dexample/exampledocs/mp500.xml -Dexample/exampledocs/sd500.xml -Dexample/exampledocs/solr.xml -Dexample/exampledocs/utf8-example.xml -Dexample/exampledocs/vidcard.xml org.apache.solr.util.SimplePostTool example/exampledocs/gb18030-example.xml
SimplePostTool version 1.5
Posting files to base url http://localhost:8983/solr/gettingstarted/update using content-type application/xml..
POSTing file gb18030-example.xml
1 files indexed.
COMMITting Solr index changes to http://localhost:8983/solr/gettingstarted/update..
Time spent: 0:00:00.074

Commit 1647928 from Erik Hatcher in branch 'dev/trunk'
[ https://svn.apache.org/r1647928 ]

SOLR-6435: Add script to simplify posting content to Solr

Put a simple stake in the ground on trunk with bin/post.

TODO's: create comparable bin/post.cmd for Windows; centralize common environment (like Java and variables) across bin/solr and bin/post; merge this to branch_5x

Commit 1648478 from Erik Hatcher in branch 'dev/trunk'
[ https://svn.apache.org/r1648478 ]

SOLR-6435: bin/post cleanup for 5x merge

Commit 1648479 from Erik Hatcher in branch 'dev/branches/branch_5x'
[ https://svn.apache.org/r1648479 ]

SOLR-6435: Added bin/post

Pleasantly functional version committed to both 5x and trunk. 'nix only at this point (no .cmd version). A brief bit of code is copied from bin/solr with TODO to centralize what makes sense across scripts.

Is 'tehfiles' intentional or misspelling?

bin/post tehfiles ~/Documents

Also. Are there any plans for something - anything - that issues a delete command?

Is 'tehfiles' intentional or misspelling?

yeah, lol - was just a comment in the script, not official usage. We'll get official -help style usage output in the script as well (feel free to open a new JIRA if you're feeling it).

Are there any plans for something - anything - that issues a delete command?

What do you have in mind? One could still do that with the SimplePostTool, with

java -Ddata=args -Dc=tehfiles -classpath dist/solr-core-*.jar org.apache.solr.util.SimplePostTool dist/solr-core-*.jar '<delete><id>42</id></delete>'

however bin/post does not (currently) support -Ddata=args. Certainly we wouldn't want a user to type in that XML incantation though, so I imagine there could be a bin/delete script that allowed for clean deleting by id or by query. I'm curious what scenarios and interface folks imagine for a friendlier delete facility.

Bulk close after 5.0 release.