You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@lucene.apache.org by kr...@apache.org on 2017/01/06 21:47:34 UTC
[14/25] lucene-solr:jira/solr-8593: SOLR-8542: README and
solr/contrib/ltr/example changes
SOLR-8542: README and solr/contrib/ltr/example changes
details:
* reduced README in favour of equivalent Solr Ref Guide content and (new) example/README
* solr/contrib/ltr/example improvements and fixes
also:
* stop supporting '*' in Managed(Feature|Model)Store.doDeleteChild
Project: http://git-wip-us.apache.org/repos/asf/lucene-solr/repo
Commit: http://git-wip-us.apache.org/repos/asf/lucene-solr/commit/eb2a8ba2
Tree: http://git-wip-us.apache.org/repos/asf/lucene-solr/tree/eb2a8ba2
Diff: http://git-wip-us.apache.org/repos/asf/lucene-solr/diff/eb2a8ba2
Branch: refs/heads/jira/solr-8593
Commit: eb2a8ba2eec0841f03bbcf7807e602f7164a606e
Parents: 2f62fac
Author: Christine Poerschke <cp...@apache.org>
Authored: Wed Dec 28 16:40:47 2016 +0000
Committer: Christine Poerschke <cp...@apache.org>
Committed: Wed Jan 4 17:12:38 2017 +0000
----------------------------------------------------------------------
solr/contrib/ltr/README.md | 390 +------------------
solr/contrib/ltr/example/README.md | 60 +++
solr/contrib/ltr/example/config.json | 13 +-
solr/contrib/ltr/example/exampleFeatures.json | 26 ++
solr/contrib/ltr/example/libsvm_formatter.py | 12 +-
.../ltr/example/techproducts-features.json | 26 --
.../contrib/ltr/example/techproducts-model.json | 18 -
.../ltr/example/train_and_upload_demo_model.py | 143 ++++---
.../ltr/store/rest/ManagedFeatureStore.java | 3 -
.../solr/ltr/store/rest/ManagedModelStore.java | 6 +-
.../multipleadditivetreesmodel.json | 2 +-
.../org/apache/solr/ltr/TestRerankBase.java | 9 +-
.../model/TestMultipleAdditiveTreesModel.java | 9 +-
.../solr/ltr/store/rest/TestModelManager.java | 47 ++-
.../store/rest/TestModelManagerPersistence.java | 4 +-
15 files changed, 237 insertions(+), 531 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/README.md
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/README.md b/solr/contrib/ltr/README.md
index 83fb279..a997d16 100644
--- a/solr/contrib/ltr/README.md
+++ b/solr/contrib/ltr/README.md
@@ -1,390 +1,24 @@
-Apache Solr Learning to Rank
+Welcome to Apache Solr Learning to Rank!
========
-This is the main [learning to rank integrated into solr](http://www.slideshare.net/lucidworks/learning-to-rank-in-solr-presented-by-michael-nilsson-diego-ceccarelli-bloomberg-lp)
-repository.
-[Read up on learning to rank](https://en.wikipedia.org/wiki/Learning_to_rank)
-
Apache Solr Learning to Rank (LTR) provides a way for you to extract features
directly inside Solr for use in training a machine learned model. You can then
deploy that model to Solr and use it to rerank your top X search results.
-# Test the plugin with solr/example/techproducts in a few easy steps!
-
-Solr provides some simple example of indices. In order to test the plugin with
-the techproducts example please follow these steps.
-If you want to install the plugin on your instance of Solr, please refer
-to the [Solr Ref Guide](https://cwiki.apache.org/confluence/display/solr/Result+Reranking).
-
-1. Compile solr and the examples
-
- `cd solr`
- `ant dist`
- `ant server`
-
-2. Run the example to setup the index, enabling the ltr plugin
-
- `./bin/solr -e techproducts -Dsolr.ltr.enabled=true`
-
-3. Deploy features and a model
-
- `curl -XPUT 'http://localhost:8983/solr/techproducts/schema/feature-store' --data-binary "@./contrib/ltr/example/techproducts-features.json" -H 'Content-type:application/json'`
-
- `curl -XPUT 'http://localhost:8983/solr/techproducts/schema/model-store' --data-binary "@./contrib/ltr/example/techproducts-model.json" -H 'Content-type:application/json'`
-
-4. Have fun !
-
- * Access to the default feature store
-
- http://localhost:8983/solr/techproducts/schema/feature-store/\_DEFAULT\_
- * Access to the model store
-
- http://localhost:8983/solr/techproducts/schema/model-store
- * Perform a reranking query using the model, and retrieve the features
-
- http://localhost:8983/solr/techproducts/query?indent=on&q=test&wt=json&rq={!ltr%20model=linear%20reRankDocs=25%20efi.user_query=%27test%27}&fl=[features],price,score,name
-
-
-BONUS: Train an actual machine learning model
-
-1. Download and install [liblinear](https://www.csie.ntu.edu.tw/~cjlin/liblinear/)
-
-2. Change `contrib/ltr/example/config.json` "trainingLibraryLocation" to point to the train directory where you installed liblinear.
-
-3. Extract features, train a reranking model, and deploy it to Solr.
-
- `cd contrib/ltr/example`
-
- `python train_and_upload_demo_model.py -c config.json`
-
- This script deploys your features from `config.json` "featuresFile" to Solr. Then it takes the relevance judged query
- document pairs of "userQueriesFile" and merges it with the features extracted from Solr into a training
- file. That file is used to train a linear model, which is then deployed to Solr for you to rerank results.
-
-4. Search and rerank the results using the trained model
-
- http://localhost:8983/solr/techproducts/query?indent=on&q=test&wt=json&rq={!ltr%20model=ExampleModel%20reRankDocs=25%20efi.user_query=%27test%27}&fl=price,score,name
-
-# Changes to solrconfig.xml
-```xml
-<config>
- ...
-
- <!-- Query parser used to rerank top docs with a provided model -->
- <queryParser name="ltr" class="org.apache.solr.ltr.search.LTRQParserPlugin" />
-
- <!-- Transformer that will encode the document features in the response.
- For each document the transformer will add the features as an extra field
- in the response. The name of the field will be the the name of the
- transformer enclosed between brackets (in this case [features]).
- In order to get the feature vector you will have to
- specify that you want the field (e.g., fl="*,[features]) -->
-
- <transformer name="features" class="org.apache.solr.ltr.response.transform.LTRFeatureLoggerTransformerFactory">
- <str name="fvCacheName">QUERY_DOC_FV</str>
- </transformer>
-
- <query>
- ...
-
- <!-- Cache for storing and fetching feature vectors -->
- <cache name="QUERY_DOC_FV"
- class="solr.search.LRUCache"
- size="4096"
- initialSize="2048"
- autowarmCount="4096"
- regenerator="solr.search.NoOpRegenerator" />
- </query>
-
-</config>
-
-```
-
-# Defining Features
-In the learning to rank plugin, you can define features in a feature space
-using standard Solr queries. As an example:
-
-###### features.json
-```json
-[
-{ "name": "isBook",
- "class": "org.apache.solr.ltr.feature.SolrFeature",
- "params":{ "fq": ["{!terms f=category}book"] }
-},
-{
- "name": "documentRecency",
- "class": "org.apache.solr.ltr.feature.SolrFeature",
- "params": {
- "q": "{!func}recip( ms(NOW,publish_date), 3.16e-11, 1, 1)"
- }
-},
-{
- "name":"originalScore",
- "class":"org.apache.solr.ltr.feature.OriginalScoreFeature",
- "params":{}
-},
-{
- "name" : "userTextTitleMatch",
- "class" : "org.apache.solr.ltr.feature.SolrFeature",
- "params" : { "q" : "{!field f=title}${user_text}" }
-},
- {
- "name" : "userFromMobile",
- "class" : "org.apache.solr.ltr.feature.ValueFeature",
- "params" : { "value" : "${userFromMobile}", "required":true }
- }
-]
-```
-
-Defines five features. Anything that is a valid Solr query can be used to define
-a feature.
-
-### Filter Query Features
-The first feature isBook fires if the term 'book' matches the category field
-for the given examined document. Since in this feature q was not specified,
-either the score 1 (in case of a match) or the score 0 (in case of no match)
-will be returned.
-
-### Query Features
-In the second feature (documentRecency) q was specified using a function query.
-In this case the score for the feature on a given document is whatever the query
-returns (1 for docs dated now, 1/2 for docs dated 1 year ago, 1/3 for docs dated
-2 years ago, etc..) . If both an fq and q is used, documents that don't match
-the fq will receive a score of 0 for the documentRecency feature, all other
-documents will receive the score specified by the query for this feature.
-
-### Original Score Feature
-The third feature (originalScore) has no parameters, and uses the
-OriginalScoreFeature class instead of the SolrFeature class. Its purpose is
-to simply return the score for the original search request against the current
-matching document.
-
-### External Features
-Users can specify external information that can to be passed in as
-part of the query to the ltr ranking framework. In this case, the
-fourth feature (userTextPhraseMatch) will be looking for an external field
-called 'user_text' passed in through the request, and will fire if there is
-a term match for the document field 'title' from the value of the external
-field 'user_text'. You can provide default values for external features as
-well by specifying ${myField:myDefault}, similar to how you would in a Solr config.
-In this case, the fifth feature (userFromMobile) will be looking for an external parameter
-called 'userFromMobile' passed in through the request, if the ValueFeature is :
-required=true, it will throw an exception if the external feature is not passed
-required=false, it will silently ignore the feature and avoid the scoring ( at Document scoring time, the model will consider 0 as feature value)
-The advantage in defining a feature as not required, where possible, is to avoid wasting caching space and time in calculating the featureScore.
-See the [Run a Rerank Query](#run-a-rerank-query) section for how to pass in external information.
-
-### Custom Features
-Custom features can be created by extending from
-org.apache.solr.ltr.feature.Feature, however this is generally not recommended.
-The majority of features should be possible to create using the methods described
-above.
-
-# Defining Models
-Currently the Learning to Rank plugin supports 2 generalized forms of
-models: 1. Linear Model i.e. [RankSVM](http://www.cs.cornell.edu/people/tj/publications/joachims_02c.pdf), [Pranking](https://papers.nips.cc/paper/2023-pranking-with-ranking.pdf)
-and 2. Multiple Additive Trees i.e. [LambdaMART](http://research.microsoft.com/pubs/132652/MSR-TR-2010-82.pdf), [Gradient Boosted Regression Trees (GBRT)](https://papers.nips.cc/paper/3305-a-general-boosting-method-and-its-application-to-learning-ranking-functions-for-web-search.pdf)
-
-### Linear
-If you'd like to introduce a bias set a constant feature
-to the bias value you'd like and make a weight of 1.0 for that feature.
-
-###### model.json
-```json
-{
- "class":"org.apache.solr.ltr.model.LinearModel",
- "name":"myModelName",
- "features":[
- { "name": "userTextTitleMatch"},
- { "name": "originalScore"},
- { "name": "isBook"}
- ],
- "params":{
- "weights": {
- "userTextTitleMatch": 1.0,
- "originalScore": 0.5,
- "isBook": 0.1
- }
-
- }
-}
-```
-
-This is an example of a toy Linear model. Class specifies the class to be
-using to interpret the model. Name is the model identifier you will use
-when making request to the ltr framework. Features specifies the feature
-space that you want extracted when using this model. All features that
-appear in the model params will be used for scoring and must appear in
-the features list. You can add extra features to the features list that
-will be computed but not used in the model for scoring, which can be useful
-for logging. Params are the Linear parameters.
-
-Good library for training SVM, an example of a Linear model, is
-(https://www.csie.ntu.edu.tw/~cjlin/liblinear/ , https://www.csie.ntu.edu.tw/~cjlin/libsvm/) .
-You will need to convert the libSVM model format to the format specified above.
-
-### Multiple Additive Trees
-
-###### model2.json
-```json
-{
- "class":"org.apache.solr.ltr.model.MultipleAdditiveTreesModel",
- "name":"multipleadditivetreesmodel",
- "features":[
- { "name": "userTextTitleMatch"},
- { "name": "originalScore"}
- ],
- "params":{
- "trees": [
- {
- "weight" : 1,
- "root": {
- "feature": "userTextTitleMatch",
- "threshold": 0.5,
- "left" : {
- "value" : -100
- },
- "right": {
- "feature" : "originalScore",
- "threshold": 10.0,
- "left" : {
- "value" : 50
- },
- "right" : {
- "value" : 75
- }
- }
- }
- },
- {
- "weight" : 2,
- "root": {
- "value" : -10
- }
- }
- ]
- }
-}
-```
-This is an example of a toy Multiple Additive Trees. Class specifies the class to be using to
-interpret the model. Name is the
-model identifier you will use when making request to the ltr framework.
-Features specifies the feature space that you want extracted when using this
-model. All features that appear in the model params will be used for scoring and
-must appear in the features list. You can add extra features to the features
-list that will be computed but not used in the model for scoring, which can
-be useful for logging. Params are the Multiple Additive Trees specific parameters. In this
-case we have 2 trees, one with 3 leaf nodes and one with 1 leaf node.
-
-A good library for training LambdaMART, an example of Multiple Additive Trees, is ( http://sourceforge.net/p/lemur/wiki/RankLib/ ).
-You will need to convert the RankLib model format to the format specified above.
-
-# Deploy Models and Features
-To send features run
-
-`curl -XPUT 'http://localhost:8983/solr/collection1/schema/feature-store' --data-binary @/path/features.json -H 'Content-type:application/json'`
-
-To send models run
-
-`curl -XPUT 'http://localhost:8983/solr/collection1/schema/model-store' --data-binary @/path/model.json -H 'Content-type:application/json'`
-
-
-# View Models and Features
-`curl -XGET 'http://localhost:8983/solr/collection1/schema/feature-store'`
-
-`curl -XGET 'http://localhost:8983/solr/collection1/schema/model-store'`
-
-# Run a Rerank Query
-Add to your original solr query
-`rq={!ltr model=myModelName reRankDocs=25}`
-
-The model name is the name of the model you sent to solr earlier.
-The number of documents you want reranked, which can be larger than the
-number you display, is reRankDocs.
-
-### Pass in external information for external features
-Add to your original solr query
-`rq={!ltr reRankDocs=3 model=externalmodel efi.field1='text1' efi.field2='text2'}`
-
-Where "field1" specifies the name of the customized field to be used by one
-or more of your features, and text1 is the information to be pass in. As an
-example that matches the earlier shown userTextTitleMatch feature one could do:
-
-`rq={!ltr reRankDocs=3 model=externalmodel efi.user_text='Casablanca' efi.user_intent='movie'}`
-
-# Extract features
-To extract features you need to use the feature vector transformer `features`
-
-`fl=*,score,[features]&rq={!ltr model=yourModel reRankDocs=25}`
-
-If you use `[features]` together with your reranking model, it will return
-the array of features used by your model. Otherwise you can just ask solr to
-produce the features without doing the reranking:
-
-`fl=*,score,[features store=yourFeatureStore format=[dense|sparse] ]`
-
-This will return the values of the features in the given store. The format of the
-extracted features will be based on the format parameter. The default is dense.
-
-# Assemble training data
-In order to train a learning to rank model you need training data. Training data is
-what "teaches" the model what the appropriate weight for each feature is. In general
-training data is a collection of queries with associated documents and what their ranking/score
-should be. As an example:
-```
-secretary of state|John Kerry|0.66|CROWDSOURCE
-secretary of state|Cesar A. Perales|0.33|CROWDSOURCE
-secretary of state|New York State|0.0|CROWDSOURCE
-secretary of state|Colorado State University Secretary|0.0|CROWDSOURCE
-
-microsoft ceo|Satya Nadella|1.0|CLICK_LOG
-microsoft ceo|Microsoft|0.0|CLICK_LOG
-microsoft ceo|State|0.0|CLICK_LOG
-microsoft ceo|Secretary|0.0|CLICK_LOG
-```
-In this example the first column indicates the query, the second column indicates a unique id for that doc,
-the third column indicates the relative importance or relevance of that doc, and the fourth column indicates the source.
-There are 2 primary ways you might collect data for use with your machine learning algorithim. The first
-is to collect the clicks of your users given a specific query. There are many ways of preparing this data
-to train a model (http://www.cs.cornell.edu/people/tj/publications/joachims_etal_05a.pdf). The general idea
-is that if a user sees multiple documents and clicks the one lower down, that document should be scored higher
-than the one above it. The second way is explicitly through a crowdsourcing platform like Mechanical Turk or
-CrowdFlower. These platforms allow you to show human workers documents associated with a query and have them
-tell you what the correct ranking should be.
+# Getting Started With Solr Learning To Rank
-At this point you'll need to collect feature vectors for each query document pair. You can use the information
-from the Extract features section above to do this. An example script has been included in example/train_and_upload_demo_model.py.
+For information on how to get started with solr ltr please see:
+[Solr Reference Guide's section on Result Reranking](https://cwiki.apache.org/confluence/display/solr/Result+Reranking)
-# Explanation of the core reranking logic
-An LTR model is plugged into the ranking through the [LTRQParserPlugin](/solr/contrib/ltr/src/java/org/apache/solr/ltr/search/LTRQParserPlugin.java). The plugin will
-read from the request the model, an instance of [LTRScoringModel](/solr/contrib/ltr/src/java/org/apache/solr/ltr/model/LTRScoringModel.java),
-plus other parameters. The plugin will generate an LTRQuery, a particular [ReRankQuery](/solr/core/src/java/org/apache/solr/search/AbstractReRankQuery.java).
-It wraps the original solr query for the first pass ranking, and uses the provided model in an
-[LTRScoringQuery](/solr/contrib/ltr/src/java/org/apache/solr/ltr/LTRScoringQuery.java) to
-rescore and rerank the top documents. The LTRScoringQuery will take care of computing the values of all the
-[features](/solr/contrib/ltr/src/java/org/apache/solr/ltr/feature/Feature.java) and then will delegate the final score
-generation to the LTRScoringModel.
+# Getting Started With Solr
-# Speeding up the weight creation with threads
-About half the time for ranking is spent in the creation of weights for each feature used in ranking. If the number of features is significantly high (say, 500 or more), this increases the ranking overhead proportionally. To alleviate this problem, parallel weight creation is provided as a configurable option. In order to use this feature, the following lines need to be added to the solrconfig.xml
-```xml
+For information on how to get started with solr please see:
+[solr/README.txt](../../README.txt)
+[Solr Quick Start](http://lucene.apache.org/solr/quickstart.html)
-<config>
- <!-- Query parser used to rerank top docs with a provided model -->
- <queryParser name="ltr" class="org.apache.solr.ltr.search.LTRQParserPlugin">
- <int name="threadModule.totalPoolThreads">10</int> <!-- Maximum threads to share for all requests -->
- <int name="threadModule.numThreadsPerRequest">5</int> <!-- Maximum threads to use for a single request -->
- </queryParser>
-
- <!-- Transformer for extracting features -->
- <transformer name="features" class="org.apache.solr.ltr.response.transform.LTRFeatureLoggerTransformerFactory">
- <str name="fvCacheName">QUERY_DOC_FV</str>
- <int name="threadModule.totalPoolThreads">10</int> <!-- Maximum threads to share for all requests -->
- <int name="threadModule.numThreadsPerRequest">5</int> <!-- Maximum threads to use for a single request -->
- </transformer>
-</config>
+# How To Contribute
-```
-
-The threadModule.totalPoolThreads option limits the total number of threads to be used across all query instances at any given time. threadModule.numThreadsPerRequest limits the number of threads used to process a single query. In the above example, 10 threads will be used to services all queries and a maximum of 5 threads to service a single query. If the solr instance is expected to receive no more than one query at a time, it is best to set both these numbers to the same value. If multiple queries need to be serviced simultaneously, the numbers can be adjusted based on the expected response times. If the value of threadModule.numThreadsPerRequest is higher, the response time for a single query will be improved upto a point. If multiple queries are serviced simultaneously, the threadModule.totalPoolThreads imposes a contention between the queries if (threadModule.numThreadsPerRequest*total parallel queries > threadModule.totalPoolThreads).
+For information on how to contribute see:
+http://wiki.apache.org/lucene-java/HowToContribute
+http://wiki.apache.org/solr/HowToContribute
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/README.md
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/README.md b/solr/contrib/ltr/example/README.md
new file mode 100644
index 0000000..1363d5d
--- /dev/null
+++ b/solr/contrib/ltr/example/README.md
@@ -0,0 +1,60 @@
+This README file is only about this example directory's content.
+
+Please refer to the Solr Reference Guide's section on [Result Reranking](https://cwiki.apache.org/confluence/display/solr/Result+Reranking) section for broader information on Learning to Rank (LTR) with Apache Solr.
+
+# Start Solr with the LTR plugin enabled
+
+ `./bin/solr -e techproducts -Dsolr.ltr.enabled=true`
+
+# Train an example machine learning model using LIBLINEAR
+
+1. Download and install [liblinear](https://www.csie.ntu.edu.tw/~cjlin/liblinear/)
+
+2. Change `contrib/ltr/example/config.json` "trainingLibraryLocation" to point to the train directory where you installed liblinear.
+
+ Alternatively, leave the `config.json` file unchanged and create a soft-link to your `liblinear` directory e.g.
+
+ `ln -s /Users/YourNameHere/Downloads/liblinear-2.1 ./contrib/ltr/example/liblinear`
+
+3. Extract features, train a reranking model, and deploy it to Solr.
+
+ `cd contrib/ltr/example`
+
+ `python train_and_upload_demo_model.py -c config.json`
+
+ This script deploys your features from `config.json` "solrFeaturesFile" to Solr. Then it takes the relevance judged query
+ document pairs of "userQueriesFile" and merges it with the features extracted from Solr into a training
+ file. That file is used to train a linear model, which is then deployed to Solr for you to rerank results.
+
+4. Search and rerank the results using the trained model
+
+ http://localhost:8983/solr/techproducts/query?indent=on&q=test&wt=json&rq={!ltr%20model=exampleModel%20reRankDocs=25%20efi.user_query=%27test%27}&fl=price,score,name
+
+# Assemble training data
+In order to train a learning to rank model you need training data. Training data is
+what "teaches" the model what the appropriate weight for each feature is. In general
+training data is a collection of queries with associated documents and what their ranking/score
+should be. As an example:
+```
+hard drive|SP2514N|0.6666666|CLICK_LOGS
+hard drive|6H500F0|0.330082034|CLICK_LOGS
+hard drive|F8V7067-APL-KIT|0.0|CLICK_LOGS
+hard drive|IW-02|0.0|CLICK_LOGS
+
+ipod|MA147LL/A|1.0|EXPLICIT
+ipod|F8V7067-APL-KIT|0.25|EXPLICIT
+ipod|IW-02|0.25|EXPLICIT
+ipod|6H500F0|0.0|EXPLICIT
+```
+In this example the first column indicates the query, the second column indicates a unique id for that doc,
+the third column indicates the relative importance or relevance of that doc, and the fourth column indicates the source.
+There are 2 primary ways you might collect data for use with your machine learning algorithim. The first
+is to collect the clicks of your users given a specific query. There are many ways of preparing this data
+to train a model (http://www.cs.cornell.edu/people/tj/publications/joachims_etal_05a.pdf). The general idea
+is that if a user sees multiple documents and clicks the one lower down, that document should be scored higher
+than the one above it. The second way is explicitly through a crowdsourcing platform like Mechanical Turk or
+CrowdFlower. These platforms allow you to show human workers documents associated with a query and have them
+tell you what the correct ranking should be.
+
+At this point you'll need to collect feature vectors for each query document pair. You can use the information
+from the Extract features section above to do this. An example script has been included in example/train_and_upload_demo_model.py.
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/config.json
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/config.json b/solr/contrib/ltr/example/config.json
index 483fe69..0b07b0e 100644
--- a/solr/contrib/ltr/example/config.json
+++ b/solr/contrib/ltr/example/config.json
@@ -4,11 +4,14 @@
"collection": "techproducts",
"requestHandler": "query",
"q": "*:*",
- "otherParams": "fl=id,score,[features efi.user_query='$USERQUERY']",
+ "efiParams": "efi.user_query='$USERQUERY'",
"userQueriesFile": "user_queries.txt",
- "trainingFile": "ClickData",
- "featuresFile": "techproducts-features.json",
+ "trainingFile": "exampleTrainingFile.txt",
+ "trainedModelFile": "exampleTrainedModel.txt",
"trainingLibraryLocation": "liblinear/train",
- "solrModelFile": "solrModel.json",
- "solrModelName": "ExampleModel"
+ "trainingLibraryOptions": "-q",
+ "solrFeaturesFile": "exampleFeatures.json",
+ "solrFeatureStoreName": "exampleFeatureStore",
+ "solrModelFile": "exampleModel.json",
+ "solrModelName": "exampleModel"
}
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/exampleFeatures.json
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/exampleFeatures.json b/solr/contrib/ltr/example/exampleFeatures.json
new file mode 100644
index 0000000..89423f3
--- /dev/null
+++ b/solr/contrib/ltr/example/exampleFeatures.json
@@ -0,0 +1,26 @@
+[
+ {
+ "store" : "exampleFeatureStore",
+ "name" : "isInStock",
+ "class" : "org.apache.solr.ltr.feature.FieldValueFeature",
+ "params" : { "field" : "inStock" }
+ },
+ {
+ "store" : "exampleFeatureStore",
+ "name" : "originalScore",
+ "class" : "org.apache.solr.ltr.feature.OriginalScoreFeature",
+ "params" : {}
+ },
+ {
+ "store" : "exampleFeatureStore",
+ "name" : "price",
+ "class" : "org.apache.solr.ltr.feature.FieldValueFeature",
+ "params" : { "field" : "price" }
+ },
+ {
+ "store" : "exampleFeatureStore",
+ "name" : "productNameMatchQuery",
+ "class" : "org.apache.solr.ltr.feature.SolrFeature",
+ "params" : { "q" : "{!field f=name}${user_query}" }
+ }
+]
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/libsvm_formatter.py
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/libsvm_formatter.py b/solr/contrib/ltr/example/libsvm_formatter.py
index 25cf10b..4bcb761 100644
--- a/solr/contrib/ltr/example/libsvm_formatter.py
+++ b/solr/contrib/ltr/example/libsvm_formatter.py
@@ -10,12 +10,12 @@ class LibSvmFormatter:
for each query in a tuple composed of: (query , docId , relevance , source , featureVector).
The list of documents that are part of the same query will generate comparisons
against each other for training. '''
- curQueryAndSource = "";
with open(trainingFile,"w") as output:
self.featureNameToId = {}
self.featureIdToName = {}
self.curFeatIndex = 1;
curListOfFv = []
+ curQueryAndSource = ""
for query,docId,relevance,source,featureVector in docClickInfo:
if curQueryAndSource != query + source:
#Time to flush out all the pairs
@@ -31,7 +31,7 @@ class LibSvmFormatter:
this requirement.'''
features = {}
for keyValuePairStr in featureVector:
- featName,featValue = keyValuePairStr.split(":");
+ featName,featValue = keyValuePairStr.split("=");
features[self._getFeatureId(featName)] = float(featValue);
return features
@@ -42,10 +42,12 @@ class LibSvmFormatter:
self.curFeatIndex += 1;
return self.featureNameToId[key];
- def convertLibSvmModelToLtrModel(self,libSvmModelLocation, outputFile, modelName):
+ def convertLibSvmModelToLtrModel(self,libSvmModelLocation,outputFile,modelName,featureStoreName):
with open(libSvmModelLocation, 'r') as inFile:
with open(outputFile,'w') as convertedOutFile:
+ # TODO: use json module instead of direct write
convertedOutFile.write('{\n\t"class":"org.apache.solr.ltr.model.LinearModel",\n')
+ convertedOutFile.write('\t"store": "' + str(featureStoreName) + '",\n')
convertedOutFile.write('\t"name": "' + str(modelName) + '",\n')
convertedOutFile.write('\t"features": [\n')
isFirst = True;
@@ -117,8 +119,8 @@ def outputLibSvmLine(sign,fvMap,outputFile):
outputFile.write(" " + str(feat) + ":" + str(fvMap[feat]));
outputFile.write("\n")
-def trainLibSvm(libraryLocation,trainingFileName):
+def trainLibSvm(libraryLocation,libraryOptions,trainingFileName,trainedModelFileName):
if os.path.isfile(libraryLocation):
- call([libraryLocation, trainingFileName])
+ call([libraryLocation, libraryOptions, trainingFileName, trainedModelFileName])
else:
raise Exception("NO LIBRARY FOUND: " + libraryLocation);
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/techproducts-features.json
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/techproducts-features.json b/solr/contrib/ltr/example/techproducts-features.json
deleted file mode 100644
index f358f8b..0000000
--- a/solr/contrib/ltr/example/techproducts-features.json
+++ /dev/null
@@ -1,26 +0,0 @@
-[
-{
- "name": "isInStock",
- "class": "org.apache.solr.ltr.feature.FieldValueFeature",
- "params": {
- "field": "inStock"
- }
-},
-{
- "name": "price",
- "class": "org.apache.solr.ltr.feature.FieldValueFeature",
- "params": {
- "field": "price"
- }
-},
-{
- "name":"originalScore",
- "class":"org.apache.solr.ltr.feature.OriginalScoreFeature",
- "params":{}
-},
-{
- "name" : "productNameMatchQuery",
- "class" : "org.apache.solr.ltr.feature.SolrFeature",
- "params" : { "q" : "{!field f=name}${user_query}" }
-}
-]
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/techproducts-model.json
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/techproducts-model.json b/solr/contrib/ltr/example/techproducts-model.json
deleted file mode 100644
index 0efded7..0000000
--- a/solr/contrib/ltr/example/techproducts-model.json
+++ /dev/null
@@ -1,18 +0,0 @@
-{
- "class":"org.apache.solr.ltr.model.LinearModel",
- "name":"linear",
- "features":[
- {"name":"isInStock"},
- {"name":"price"},
- {"name":"originalScore"},
- {"name":"productNameMatchQuery"}
- ],
- "params":{
- "weights":{
- "isInStock":15.0,
- "price":1.0,
- "originalScore":5.0,
- "productNameMatchQuery":1.0
- }
- }
-}
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/example/train_and_upload_demo_model.py
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/example/train_and_upload_demo_model.py b/solr/contrib/ltr/example/train_and_upload_demo_model.py
index c3762de..3862456 100755
--- a/solr/contrib/ltr/example/train_and_upload_demo_model.py
+++ b/solr/contrib/ltr/example/train_and_upload_demo_model.py
@@ -10,22 +10,61 @@ from optparse import OptionParser
solrQueryUrl = ""
-def generateQueries(config):
- with open(config["userQueriesFile"]) as input:
+
+def setupSolr(collection, host, port, featuresFile, featureStoreName):
+ '''Sets up solr with the proper features for the test'''
+
+ conn = httplib.HTTPConnection(host, port)
+
+ baseUrl = "/solr/" + collection
+ featureUrl = baseUrl + "/schema/feature-store"
+
+ conn.request("DELETE", featureUrl+"/"+featureStoreName)
+ r = conn.getresponse()
+ msg = r.read()
+ if (r.status != httplib.OK and
+ r.status != httplib.CREATED and
+ r.status != httplib.ACCEPTED and
+ r.status != httplib.NOT_FOUND):
+ raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
+
+
+ # Add features
+ headers = {'Content-type': 'application/json'}
+ featuresBody = open(featuresFile)
+
+ conn.request("POST", featureUrl, featuresBody, headers)
+ r = conn.getresponse()
+ msg = r.read()
+ if (r.status != httplib.OK and
+ r.status != httplib.ACCEPTED):
+ print r.status
+ print ""
+ print r.reason;
+ raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
+
+ conn.close()
+
+
+def generateQueries(userQueriesFile, collection, requestHandler, solrFeatureStoreName, efiParams):
+ with open(userQueriesFile) as input:
solrQueryUrls = [] #A list of tuples with solrQueryUrl,solrQuery,docId,scoreForPQ,source
for line in input:
line = line.strip();
searchText,docId,score,source = line.split("|");
- solrQuery = generateHttpRequest(config,searchText,docId)
+ solrQuery = generateHttpRequest(collection,requestHandler,solrFeatureStoreName,efiParams,searchText,docId)
solrQueryUrls.append((solrQuery,searchText,docId,score,source))
return solrQueryUrls;
-def generateHttpRequest(config,searchText,docId):
+
+def generateHttpRequest(collection, requestHandler, solrFeatureStoreName, efiParams, searchText, docId):
global solrQueryUrl
if len(solrQueryUrl) < 1:
- solrQueryUrl = "/solr/%(collection)s/%(requestHandler)s?%(otherParams)s&q=" % config
+ solrQueryUrl = "/".join([ "", "solr", collection, requestHandler ])
+ solrQueryUrl += ("?fl=" + ",".join([ "id", "score", "[features store="+solrFeatureStoreName+" "+efiParams+"]" ]))
+ solrQueryUrl += "&q="
solrQueryUrl = solrQueryUrl.replace(" ","+")
solrQueryUrl += urllib.quote_plus("id:")
@@ -36,10 +75,11 @@ def generateHttpRequest(config,searchText,docId):
return solrQuery
-def generateTrainingData(solrQueries, config):
+
+def generateTrainingData(solrQueries, host, port):
'''Given a list of solr queries, yields a tuple of query , docId , score , source , feature vector for each query.
- Feature Vector is a list of strings of form "key:value"'''
- conn = httplib.HTTPConnection(config["host"], config["port"])
+ Feature Vector is a list of strings of form "key=value"'''
+ conn = httplib.HTTPConnection(host, port)
headers = {"Connection":" keep-alive"}
try:
@@ -64,7 +104,7 @@ def generateTrainingData(solrQueries, config):
if r.status == httplib.OK:
#print "http connection was ok for: " + queryUrl
- yield(query,docId,score,source,fv.split(";"));
+ yield(query,docId,score,source,fv.split(","));
else:
raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
except Exception as e:
@@ -73,40 +113,29 @@ def generateTrainingData(solrQueries, config):
conn.close()
-def setupSolr(config):
- '''Sets up solr with the proper features for the test'''
-
- conn = httplib.HTTPConnection(config["host"], config["port"])
-
- baseUrl = "/solr/" + config["collection"]
- featureUrl = baseUrl + "/schema/feature-store"
-
- # CAUTION! This will delete all feature stores. This is just for demo purposes
- conn.request("DELETE", featureUrl+"/*")
- r = conn.getresponse()
- msg = r.read()
- if (r.status != httplib.OK and
- r.status != httplib.CREATED and
- r.status != httplib.ACCEPTED and
- r.status != httplib.NOT_FOUND):
- raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
-
- # Add features
+def uploadModel(collection, host, port, modelFile, modelName):
+ modelUrl = "/solr/" + collection + "/schema/model-store"
headers = {'Content-type': 'application/json'}
- featuresBody = open(config["featuresFile"])
+ with open(modelFile) as modelBody:
+ conn = httplib.HTTPConnection(host, port)
- conn.request("POST", featureUrl, featuresBody, headers)
- r = conn.getresponse()
- msg = r.read()
- if (r.status != httplib.OK and
- r.status != httplib.ACCEPTED):
- print r.status
- print ""
- print r.reason;
- raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
+ conn.request("DELETE", modelUrl+"/"+modelName)
+ r = conn.getresponse()
+ msg = r.read()
+ if (r.status != httplib.OK and
+ r.status != httplib.CREATED and
+ r.status != httplib.ACCEPTED and
+ r.status != httplib.NOT_FOUND):
+ raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
- conn.close()
+ conn.request("POST", modelUrl, modelBody, headers)
+ r = conn.getresponse()
+ msg = r.read()
+ if (r.status != httplib.OK and
+ r.status != httplib.CREATED and
+ r.status != httplib.ACCEPTED):
+ raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
def main(argv=None):
@@ -126,38 +155,26 @@ def main(argv=None):
with open(options.configFile) as configFile:
config = json.load(configFile)
- print "Uploading feature space to Solr"
- setupSolr(config)
+ print "Uploading features ("+config["solrFeaturesFile"]+") to Solr"
+ setupSolr(config["collection"], config["host"], config["port"], config["solrFeaturesFile"], config["solrFeatureStoreName"])
- print "Generating feature extraction Solr queries"
- reRankQueries = generateQueries(config)
+ print "Converting user queries ("+config["userQueriesFile"]+") into Solr queries for feature extraction"
+ reRankQueries = generateQueries(config["userQueriesFile"], config["collection"], config["requestHandler"], config["solrFeatureStoreName"], config["efiParams"])
- print "Extracting features"
- fvGenerator = generateTrainingData(reRankQueries, config);
+ print "Running Solr queries to extract features"
+ fvGenerator = generateTrainingData(reRankQueries, config["host"], config["port"])
formatter = libsvm_formatter.LibSvmFormatter();
formatter.processQueryDocFeatureVector(fvGenerator,config["trainingFile"]);
- print "Training ranksvm model"
- libsvm_formatter.trainLibSvm(config["trainingLibraryLocation"],config["trainingFile"])
+ print "Training model using '"+config["trainingLibraryLocation"]+" "+config["trainingLibraryOptions"]+"'"
+ libsvm_formatter.trainLibSvm(config["trainingLibraryLocation"],config["trainingLibraryOptions"],config["trainingFile"],config["trainedModelFile"])
- print "Converting ranksvm model to solr model"
- formatter.convertLibSvmModelToLtrModel(config["trainingFile"] + ".model", config["solrModelFile"], config["solrModelName"])
+ print "Converting trained model ("+config["trainedModelFile"]+") to solr model ("+config["solrModelFile"]+")"
+ formatter.convertLibSvmModelToLtrModel(config["trainedModelFile"], config["solrModelFile"], config["solrModelName"], config["solrFeatureStoreName"])
- print "Uploading model to solr"
- uploadModel(config["collection"], config["host"], config["port"], config["solrModelFile"])
+ print "Uploading model ("+config["solrModelFile"]+") to Solr"
+ uploadModel(config["collection"], config["host"], config["port"], config["solrModelFile"], config["solrModelName"])
-def uploadModel(collection, host, port, modelFile):
- modelUrl = "/solr/" + collection + "/schema/model-store"
- headers = {'Content-type': 'application/json'}
- with open(modelFile) as modelBody:
- conn = httplib.HTTPConnection(host, port)
- conn.request("POST", modelUrl, modelBody, headers)
- r = conn.getresponse()
- msg = r.read()
- if (r.status != httplib.OK and
- r.status != httplib.CREATED and
- r.status != httplib.ACCEPTED):
- raise Exception("Status: {0} {1}\nResponse: {2}".format(r.status, r.reason, msg))
if __name__ == '__main__':
sys.exit(main())
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedFeatureStore.java
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedFeatureStore.java b/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedFeatureStore.java
index 2c7bce5..6b415a6 100644
--- a/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedFeatureStore.java
+++ b/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedFeatureStore.java
@@ -149,9 +149,6 @@ public class ManagedFeatureStore extends ManagedResource implements ManagedResou
@Override
public synchronized void doDeleteChild(BaseSolrResource endpoint, String childId) {
- if (childId.equals("*")) {
- stores.clear();
- }
if (stores.containsKey(childId)) {
stores.remove(childId);
}
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedModelStore.java
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedModelStore.java b/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedModelStore.java
index 9c19b0a..79640c1 100644
--- a/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedModelStore.java
+++ b/solr/contrib/ltr/src/java/org/apache/solr/ltr/store/rest/ManagedModelStore.java
@@ -167,11 +167,7 @@ public class ManagedModelStore extends ManagedResource implements ManagedResourc
@Override
public synchronized void doDeleteChild(BaseSolrResource endpoint, String childId) {
- if (childId.equals("*")) {
- store.clear();
- } else {
- store.delete(childId);
- }
+ store.delete(childId);
storeManagedData(applyUpdatesToManagedData(null));
}
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/test-files/modelExamples/multipleadditivetreesmodel.json
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/test-files/modelExamples/multipleadditivetreesmodel.json b/solr/contrib/ltr/src/test-files/modelExamples/multipleadditivetreesmodel.json
index 37551a0..70c5563 100644
--- a/solr/contrib/ltr/src/test-files/modelExamples/multipleadditivetreesmodel.json
+++ b/solr/contrib/ltr/src/test-files/modelExamples/multipleadditivetreesmodel.json
@@ -16,7 +16,7 @@
"value" : "-100"
},
"right": {
- "feature" : "this_feature_doesnt_exist",
+ "feature" : "constantScoreToForceMultipleAdditiveTreesScoreAllDocs",
"threshold": "10.0f",
"left" : {
"value" : "50"
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/test/org/apache/solr/ltr/TestRerankBase.java
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/test/org/apache/solr/ltr/TestRerankBase.java b/solr/contrib/ltr/src/test/org/apache/solr/ltr/TestRerankBase.java
index 5277821..6f5e0ce 100644
--- a/solr/contrib/ltr/src/test/org/apache/solr/ltr/TestRerankBase.java
+++ b/solr/contrib/ltr/src/test/org/apache/solr/ltr/TestRerankBase.java
@@ -43,6 +43,7 @@ import org.apache.solr.ltr.feature.ValueFeature;
import org.apache.solr.ltr.model.LTRScoringModel;
import org.apache.solr.ltr.model.LinearModel;
import org.apache.solr.ltr.model.ModelException;
+import org.apache.solr.ltr.store.FeatureStore;
import org.apache.solr.ltr.store.rest.ManagedFeatureStore;
import org.apache.solr.ltr.store.rest.ManagedModelStore;
import org.apache.solr.request.SolrQueryRequestBase;
@@ -311,6 +312,12 @@ public class TestRerankBase extends RestTestBase {
public static LTRScoringModel createModelFromFiles(String modelFileName,
String featureFileName) throws ModelException, Exception {
+ return createModelFromFiles(modelFileName, featureFileName,
+ FeatureStore.DEFAULT_FEATURE_STORE_NAME);
+ }
+
+ public static LTRScoringModel createModelFromFiles(String modelFileName,
+ String featureFileName, String featureStoreName) throws ModelException, Exception {
URL url = TestRerankBase.class.getResource("/modelExamples/"
+ modelFileName);
final String modelJson = FileUtils.readFileToString(new File(url.toURI()),
@@ -331,7 +338,7 @@ public class TestRerankBase extends RestTestBase {
final ManagedFeatureStore fs = getManagedFeatureStore();
// fs.getFeatureStore(null).clear();
- fs.doDeleteChild(null, "*"); // is this safe??
+ fs.doDeleteChild(null, featureStoreName); // is this safe??
// based on my need to call this I dont think that
// "getNewManagedFeatureStore()"
// is actually returning a new feature store each time
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/test/org/apache/solr/ltr/model/TestMultipleAdditiveTreesModel.java
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/test/org/apache/solr/ltr/model/TestMultipleAdditiveTreesModel.java b/solr/contrib/ltr/src/test/org/apache/solr/ltr/model/TestMultipleAdditiveTreesModel.java
index 1824eb4..3c817d2 100644
--- a/solr/contrib/ltr/src/test/org/apache/solr/ltr/model/TestMultipleAdditiveTreesModel.java
+++ b/solr/contrib/ltr/src/test/org/apache/solr/ltr/model/TestMultipleAdditiveTreesModel.java
@@ -38,10 +38,6 @@ public class TestMultipleAdditiveTreesModel extends TestRerankBase {
assertU(adoc("id", "4", "title", "w4", "description", "w4", "popularity","4"));
assertU(adoc("id", "5", "title", "w5", "description", "w5", "popularity","5"));
assertU(commit());
-
- loadFeatures("multipleadditivetreesmodel_features.json"); // currently needed to force
- // scoring on all docs
- loadModels("multipleadditivetreesmodel.json");
}
@AfterClass
@@ -52,6 +48,9 @@ public class TestMultipleAdditiveTreesModel extends TestRerankBase {
@Test
public void testMultipleAdditiveTreesScoringWithAndWithoutEfiFeatureMatches() throws Exception {
+ loadFeatures("multipleadditivetreesmodel_features.json");
+ loadModels("multipleadditivetreesmodel.json");
+
final SolrQuery query = new SolrQuery();
query.setQuery("*:*");
query.add("rows", "3");
@@ -75,7 +74,7 @@ public class TestMultipleAdditiveTreesModel extends TestRerankBase {
query.add("rq", "{!ltr reRankDocs=3 model=multipleadditivetreesmodel efi.user_query=w3}");
assertJQ("/query" + query.toQueryString(), "/response/docs/[0]/id=='3'");
- assertJQ("/query" + query.toQueryString(), "/response/docs/[0]/score==-20.0");
+ assertJQ("/query" + query.toQueryString(), "/response/docs/[0]/score==30.0");
assertJQ("/query" + query.toQueryString(), "/response/docs/[1]/score==-120.0");
assertJQ("/query" + query.toQueryString(), "/response/docs/[2]/score==-120.0");
}
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManager.java
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManager.java b/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManager.java
index 5b43c60..125f452 100644
--- a/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManager.java
+++ b/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManager.java
@@ -23,10 +23,10 @@ import org.apache.solr.ltr.feature.FieldValueFeature;
import org.apache.solr.ltr.feature.ValueFeature;
import org.apache.solr.ltr.model.LinearModel;
import org.apache.solr.ltr.search.LTRQParserPlugin;
+import org.apache.solr.ltr.store.FeatureStore;
import org.apache.solr.rest.ManagedResource;
import org.apache.solr.rest.ManagedResourceStorage;
import org.apache.solr.rest.RestManager;
-import org.junit.Before;
import org.junit.BeforeClass;
import org.junit.Test;
@@ -37,13 +37,6 @@ public class TestModelManager extends TestRerankBase {
setuptest(true);
}
- @Before
- public void restart() throws Exception {
- restTestHarness.delete(ManagedFeatureStore.REST_END_POINT + "/*");
- restTestHarness.delete(ManagedModelStore.REST_END_POINT + "/*");
-
- }
-
@Test
public void test() throws Exception {
final SolrResourceLoader loader = new SolrResourceLoader(
@@ -76,6 +69,7 @@ public class TestModelManager extends TestRerankBase {
@Test
public void testRestManagerEndpoints() throws Exception {
+ final String TEST_FEATURE_STORE_NAME = "TEST";
// relies on these ManagedResources being activated in the
// schema-rest.xml used by this test
assertJQ("/schema/managed", "/responseHeader/status==0");
@@ -95,7 +89,7 @@ public class TestModelManager extends TestRerankBase {
assertJPut(ManagedFeatureStore.REST_END_POINT, feature,
"/responseHeader/status==0");
- feature = "{\"name\": \"test33\", \"store\": \"TEST\", \"class\": \""+valueFeatureClassName+"\", \"params\": {\"value\": 1} }";
+ feature = "{\"name\": \"test33\", \"store\": \""+TEST_FEATURE_STORE_NAME+"\", \"class\": \""+valueFeatureClassName+"\", \"params\": {\"value\": 1} }";
assertJPut(ManagedFeatureStore.REST_END_POINT, feature,
"/responseHeader/status==0");
@@ -136,17 +130,26 @@ public class TestModelManager extends TestRerankBase {
assert (qryResult.contains("\"name\":\"testmodel3\"")
&& qryResult.contains("\"name\":\"testmodel4\"") && qryResult
.contains("\"name\":\"testmodel5\""));
- /*
- * assertJQ(LTRParams.MSTORE_END_POINT, "/models/[0]/name=='testmodel3'");
- * assertJQ(LTRParams.MSTORE_END_POINT, "/models/[1]/name=='testmodel4'");
- * assertJQ(LTRParams.MSTORE_END_POINT, "/models/[2]/name=='testmodel5'");
- */
+
+ assertJQ(ManagedModelStore.REST_END_POINT, "/models/[0]/name=='testmodel3'");
+ assertJQ(ManagedModelStore.REST_END_POINT, "/models/[1]/name=='testmodel4'");
+ assertJQ(ManagedModelStore.REST_END_POINT, "/models/[2]/name=='testmodel5'");
+ restTestHarness.delete(ManagedModelStore.REST_END_POINT + "/testmodel3");
+ restTestHarness.delete(ManagedModelStore.REST_END_POINT + "/testmodel4");
+ restTestHarness.delete(ManagedModelStore.REST_END_POINT + "/testmodel5");
+ assertJQ(ManagedModelStore.REST_END_POINT + "/" + FeatureStore.DEFAULT_FEATURE_STORE_NAME,
+ "/models==[]'");
+
assertJQ(ManagedFeatureStore.REST_END_POINT,
- "/featureStores==['TEST','_DEFAULT_']");
- assertJQ(ManagedFeatureStore.REST_END_POINT + "/_DEFAULT_",
+ "/featureStores==['"+TEST_FEATURE_STORE_NAME+"','"+FeatureStore.DEFAULT_FEATURE_STORE_NAME+"']");
+ assertJQ(ManagedFeatureStore.REST_END_POINT + "/" + FeatureStore.DEFAULT_FEATURE_STORE_NAME,
"/features/[0]/name=='test1'");
- assertJQ(ManagedFeatureStore.REST_END_POINT + "/TEST",
+ assertJQ(ManagedFeatureStore.REST_END_POINT + "/"+TEST_FEATURE_STORE_NAME,
"/features/[0]/name=='test33'");
+ restTestHarness.delete(ManagedFeatureStore.REST_END_POINT + "/" + FeatureStore.DEFAULT_FEATURE_STORE_NAME);
+ restTestHarness.delete(ManagedFeatureStore.REST_END_POINT + "/"+TEST_FEATURE_STORE_NAME);
+ assertJQ(ManagedFeatureStore.REST_END_POINT,
+ "/featureStores==[]");
}
@Test
@@ -154,10 +157,16 @@ public class TestModelManager extends TestRerankBase {
loadFeatures("features-linear.json");
loadModels("linear-model.json");
+ final String modelName = "6029760550880411648";
assertJQ(ManagedModelStore.REST_END_POINT,
- "/models/[0]/name=='6029760550880411648'");
- assertJQ(ManagedFeatureStore.REST_END_POINT + "/_DEFAULT_",
+ "/models/[0]/name=='"+modelName+"'");
+ assertJQ(ManagedFeatureStore.REST_END_POINT + "/" + FeatureStore.DEFAULT_FEATURE_STORE_NAME,
+ "/features/[0]/name=='title'");
+ assertJQ(ManagedFeatureStore.REST_END_POINT + "/" + FeatureStore.DEFAULT_FEATURE_STORE_NAME,
"/features/[1]/name=='description'");
+
+ restTestHarness.delete(ManagedModelStore.REST_END_POINT + "/"+modelName);
+ restTestHarness.delete(ManagedFeatureStore.REST_END_POINT + "/" + FeatureStore.DEFAULT_FEATURE_STORE_NAME);
}
}
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/eb2a8ba2/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManagerPersistence.java
----------------------------------------------------------------------
diff --git a/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManagerPersistence.java b/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManagerPersistence.java
index f707e55..9168dd0 100644
--- a/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManagerPersistence.java
+++ b/solr/contrib/ltr/src/test/org/apache/solr/ltr/store/rest/TestModelManagerPersistence.java
@@ -106,9 +106,9 @@ public class TestModelManagerPersistence extends TestRerankBase {
assertJQ(ManagedModelStore.REST_END_POINT + "/test-model2",
"/models/[0]/name=='test-model'");
- assertJDelete(ManagedModelStore.REST_END_POINT + "/*",
+ assertJDelete(ManagedModelStore.REST_END_POINT + "/test-model1",
"/responseHeader/status==0");
- assertJDelete(ManagedFeatureStore.REST_END_POINT + "/*",
+ assertJDelete(ManagedFeatureStore.REST_END_POINT + "/test1",
"/responseHeader/status==0");
assertJQ(ManagedFeatureStore.REST_END_POINT + "/test1",
"/features/==[]");