Re: New Committer: Nikhil Kak
Congrats Nikhil! On Wed, Jun 27, 2018 at 1:39 PM, Nandish Jayaram wrote: > Dear MADlib dev community, > > The Project Management Committee (PMC) for Apache MADlib > has invited Nikhil to become a committer and we are pleased > to announce that he has accepted. > > Nikhil started working on the project in Nov 2017. Since that > time, he has contributed significantly in both features and bug > fixes in the following areas: > > - mini-batch preprocessor > - utilities > - neural networks/multilayer perceptron > - correlation > - HITS graph algorithm > - support vector machines > - LDA > - infrastructure projects > - documentation > - testing > > Being a committer enables easier contribution to the > project since there is no need to go via the patch > submission process. This should enable better productivity. > > Welcome Nikhil! > > Regards, > The Apache MADlib PMC >
[GitHub] madlib issue #284: SVM: Fix flaky dev-check failure
Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/284 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/528/ ---
[GitHub] madlib issue #284: SVM: Fix flaky dev-check failure
Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/284 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/527/ ---
[GitHub] madlib issue #284: SVM: Fix flaky dev-check failure
Github user njayaram2 commented on the issue: https://github.com/apache/madlib/pull/284 Thank you for the comment @iyerr3 , will relax the constraint as suggested. ---
[GitHub] madlib issue #284: SVM: Fix flaky dev-check failure
Github user iyerr3 commented on the issue: https://github.com/apache/madlib/pull/284 This test isn't necessarily the best thing to check but if we do want to keep it with some relaxed constraints then maybe we allow the norm to go higher by a small amount iff the total loss is lower. So something along the lines of ``` norm2(l2.coef) < norm2(noreg.coef) OR ( (norm2(l2.coef)-norm2(noreg.coef))/norm2(noreg.coef) < 0.1 AND l2.loss < noreg.loss), ``` ---
[GitHub] madlib issue #283: Bugfix: Fix failing dev check in CRF
Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/283 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/526/ ---
[GitHub] madlib pull request #284: SVM: Fix flaky dev-check failure
GitHub user njayaram2 opened a pull request: https://github.com/apache/madlib/pull/284 SVM: Fix flaky dev-check failure JIRA: MADLIB-1232 SVM has a dev-check query that is flaky on a large cluster. This commit relaxes the assert condition for that query. Closes #284 You can merge this pull request into a Git repository by running: $ git pull https://github.com/madlib/madlib bugfix/svm-flaky-dev-check Alternatively you can review and apply these changes as the patch at: https://github.com/apache/madlib/pull/284.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #284 commit 9c556816bcca990e9b168cf556ce0da0cacf935a Author: Nandish Jayaram Date: 2018-06-27T19:40:19Z SVM: Fix flaky dev-check failure JIRA: MADLIB-1232 SVM has a dev-check query that is flaky on a large cluster. This commit relaxes the assert condition for that query. Closes #284 ---
[GitHub] madlib pull request #283: Bugfix: Fix failing dev check in CRF
Github user asfgit closed the pull request at: https://github.com/apache/madlib/pull/283 ---
[GitHub] madlib issue #283: Bugfix: Fix failing dev check in CRF
Github user njayaram2 commented on the issue: https://github.com/apache/madlib/pull/283 Thank you for the comments @kaknikhil . I moved out the jenkins build script to a different commit. ---
[GitHub] madlib pull request #283: Bugfix: Fix failing dev check in CRF
Github user kaknikhil commented on a diff in the pull request:
https://github.com/apache/madlib/pull/283#discussion_r198641216
--- Diff: src/ports/postgres/modules/crf/test/crf_train_large.sql_in ---
@@ -234,26 +234,40 @@ INSERT INTO train_new_segmenttbl VALUES
(30, 7, 'years', 13, 31),
(31, 7, '.', 44, 31);
-CREATE TABLE train_new_regex(pattern text,name text);
+CREATE TABLE train_new_regex(pattern text,name text);
INSERT INTO train_new_regex VALUES
-('^[A-Z][a-z]+$','InitCapital'), ('^[A-Z]+$','isAllCapital'),
+('^[A-Z][a-z]+$','InitCapital'), ('^[A-Z]+$','isAllCapital'),
('^.*[0-9]+.*$','containsDigit'),('^.+[.]$','endsWithDot'),
('^.+[,]$','endsWithComma'), ('^.+er$','endsWithER'),
('^.+est$','endsWithEst'), ('^.+ed$','endsWithED'),
('^.+s$','endsWithS'), ('^.+ing$','endsWithIng'),
('^.+ly$','endsWithly'),
('^.+-.+$','isDashSeparatedWords'),
('^.*@.*$','isEmailId');
-analyze train_new_regex;
+analyze train_new_regex;
-SELECT crf_train_fgen('train_new_segmenttbl', 'train_new_regex',
'crf_label', 'train_new_dictionary',
'train_new_featuretbl','train_new_featureset');
+CREATE TABLE crf_label_new (id integer,label character varying);
--- End diff --
The two files `crf_test_small.sql_in` and `crf_train_large.sql_in` have
different indentation. Can we make them consistent
---
[GitHub] madlib pull request #283: Bugfix: Fix failing dev check in CRF
Github user kaknikhil commented on a diff in the pull request:
https://github.com/apache/madlib/pull/283#discussion_r198644242
--- Diff: src/ports/postgres/modules/crf/test/crf_test_small.sql_in ---
@@ -90,7 +90,7 @@
(18,'PRP$'),(19,'RB'), (20,'RBR'), (21,'RBS'), (22,'RP'),
(23,'SYM'), (24,'TO'), (25,'UH'), (26,'VB'),
(27,'VBD'), (28,'VBG'),(29,'VBN'), (30,'VBP'),
(31,'VBZ'),(32,'WDT'), (33,'WP'), (34,'WP$'),(35,'WRB'),
(36,'$'), (37,'#'), (38,''), (39,'``'), (40,'('),
(41,')'), (42,','), (43,'.'), (44,':');
- analyze crf_label;
+ analyze test_crf_label;
--- End diff --
Assuming that the table `crf_label` doesn't exist, why wasn't crf install
check always red?
---
New Committer: Nikhil Kak
Dear MADlib dev community, The Project Management Committee (PMC) for Apache MADlib has invited Nikhil to become a committer and we are pleased to announce that he has accepted. Nikhil started working on the project in Nov 2017. Since that time, he has contributed significantly in both features and bug fixes in the following areas: - mini-batch preprocessor - utilities - neural networks/multilayer perceptron - correlation - HITS graph algorithm - support vector machines - LDA - infrastructure projects - documentation - testing Being a committer enables easier contribution to the project since there is no need to go via the patch submission process. This should enable better productivity. Welcome Nikhil! Regards, The Apache MADlib PMC
[GitHub] madlib issue #283: Bugfix: Fix failing dev check in CRF
Github user asfgit commented on the issue: https://github.com/apache/madlib/pull/283 Refer to this link for build results (access rights to CI server needed): https://builds.apache.org/job/madlib-pr-build/525/ ---
[GitHub] madlib pull request #283: Bugfix: Fix failing dev check in CRF
GitHub user njayaram2 opened a pull request: https://github.com/apache/madlib/pull/283 Bugfix: Fix failing dev check in CRF This commit has the following changes: - A couple of dev check files in CRF did not have the label table creation in it. But the label table was consumed by one of the queries that led to dev-check failure. - Run dev check on Jenkins build instead of install check. You can merge this pull request into a Git repository by running: $ git pull https://github.com/madlib/madlib bugfix/crf-dev-check Alternatively you can review and apply these changes as the patch at: https://github.com/apache/madlib/pull/283.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #283 commit deb206b7d1c1e7ce87d6e33c7a1dff91b3adb82b Author: Nandish Jayaram Date: 2018-06-27T18:25:46Z Bugfix: Fix failing dev check in CRF This commit has the following changes: - A couple of dev check files in CRF did not have the label table creation in it. But the label table was consumed by one of the queries that led to dev-check failure. - Run dev check on Jenkins build instead of install check. ---
[GitHub] madlib pull request #282: Utilites: Add CTAS while dropping some columns
GitHub user iyerr3 opened a pull request: https://github.com/apache/madlib/pull/282 Utilites: Add CTAS while dropping some columns JIRA: MADLIB-1241 This commit adds function to create a new table from existing table while dropping some of the columns of the source table. Closes #282 You can merge this pull request into a Git repository by running: $ git pull https://github.com/madlib/madlib feature/drop_columns Alternatively you can review and apply these changes as the patch at: https://github.com/apache/madlib/pull/282.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #282 commit 3b5310c07c2092059cd116c82457d026f968cd7a Author: Rahul Iyer Date: 2018-06-27T00:36:39Z Utilites: Add CTAS while dropping some columns JIRA: MADLIB-1241 This commit adds function to create a new table from existing table while dropping some of the columns of the source table. Closes #282 ---
