You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@atlas.apache.org by Ashutosh Mestry <am...@hortonworks.com> on 2017/03/15 23:04:52 UTC

Review Request 57649: Export API: ZIP File Size Optimization

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/
-----------------------------------------------------------

Review request for atlas and Madhan Neethiraj.


Bugs: ATLAS-1503
    https://issues.apache.org/jira/browse/ATLAS-1503


Repository: atlas


Description
-------

**Background**
==============
Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.

**Solution**
============
The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)

This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.

**Implementation Details**
==========================
*Export API*
- Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
- _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
- _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.

*Import API*
- _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
- _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
- _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
- _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.


**Functional Areas Impacted**
=============================
*Export*
- Full
- Connected
- HDFS path-based import.

*Import*
- Regular flow.

**Examples**
============
Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.

Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.

**Metrics**
===========

Date | File Size | No. of Entities | Export   |
     |           |                 | Duration |
-----|-----------|-----------------|----------|
3/02 |    180 MB |          202930 |   22 mins|
3/08 |      7 KB |               3 |    5 secs|
----------------------------------------------|
               Improvement                    |           
----------------------------------------------|
3/14 |     38 MB |          202930 |   19 mins|
3/14 |      5 KB |               3 |    5 secs|


**Summary**
===========
With these changes the file size reduction is: ~65%.


Diffs
-----

  intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
  intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityImportStream.java 73994b9 
  repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
  webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
  webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 


Diff: https://reviews.apache.org/r/57649/diff/1/


Testing
-------

Test data:
- QuickStart_v1: 3 databases.
- A *hive_db* with 922 tables.
- Stocks *hive_db* with 1 database, table, process and 5 columns.
- A *hive_db* with 522K entities.

The changes impact all the flows in the Export & Import APIs.
Unit testing: Manual.
Integration testing: Manual.
Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.


Thanks,

Ashutosh Mestry


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Madhan Neethiraj <ma...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/#review169083
-----------------------------------------------------------




repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java
Lines 31 (patched)
<https://reviews.apache.org/r/57649/#comment241400>

    Instead of adding nextWithExtInfo(), consider updating AtlasEntityStream.next() to return AtlasEntityWithExtInfo - as shown below:
    
    class AtlasEntityStream {
      public AtlasEntityWithExtInfo getNext() {
        return iterator.hasNext() ? new AtlasEntityWithExtInfo(iterator.next(), this.entitiesWithExtInfo) : null; 
      }
    }
    
    With this change, following methods can be removed:
     AtlasEntityStreamForImport.nextWithExtInfo()
     AtlasEntityStreamForImport.getByGuid()
     EntityImportStream.nextWithExtInfo()



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Line 224 (original), 215 (patched)
<https://reviews.apache.org/r/57649/#comment241402>

    entityWithExtInfo.getReferredEntities() - this could be null. Please review all usage and handle this case.



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Lines 294 (patched)
<https://reviews.apache.org/r/57649/#comment241403>

    Consider sending direction as a parameter to addToBeProcessed(guid, isLineage, direction) and remove line #293, #297, #343.
    
    addToBeProcessed() can update the directory when isLineage=true, if needed.



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Line 398 (original), 395 (patched)
<https://reviews.apache.org/r/57649/#comment241404>

    "Object" ==> "Boolean"?



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Lines 439 (patched)
<https://reviews.apache.org/r/57649/#comment241405>

    Consider renaming: "ListOptmizedForContains" ==> "UniqueList"



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Lines 463 (patched)
<https://reviews.apache.org/r/57649/#comment241406>

    list.addAll() may end up adding duplicate items to the list. Consider iterating 's' and add only elements that are not present in 'set'


- Madhan Neethiraj


On March 15, 2017, 11:04 p.m., Ashutosh Mestry wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57649/
> -----------------------------------------------------------
> 
> (Updated March 15, 2017, 11:04 p.m.)
> 
> 
> Review request for atlas and Madhan Neethiraj.
> 
> 
> Bugs: ATLAS-1503
>     https://issues.apache.org/jira/browse/ATLAS-1503
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> **Background**
> ==============
> Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.
> 
> **Solution**
> ============
> The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)
> 
> This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.
> 
> **Implementation Details**
> ==========================
> *Export API*
> - Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
> - _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
> - _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.
> 
> *Import API*
> - _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
> - _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
> - _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
> - _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.
> 
> 
> **Functional Areas Impacted**
> =============================
> *Export*
> - Full
> - Connected
> - HDFS path-based import.
> 
> *Import*
> - Regular flow.
> 
> **Examples**
> ============
> Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.
> 
> Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.
> 
> **Metrics**
> ===========
> 
> Date | File Size | No. of Entities | Export   |
>      |           |                 | Duration |
> -----|-----------|-----------------|----------|
> 3/02 |    180 MB |          202930 |   22 mins|
> 3/08 |      7 KB |               3 |    5 secs|
> ----------------------------------------------|
>                Improvement                    |           
> ----------------------------------------------|
> 3/14 |     38 MB |          202930 |   19 mins|
> 3/14 |      5 KB |               3 |    5 secs|
> 
> 
> **Summary**
> ===========
> With these changes the file size reduction is: ~65%.
> 
> 
> Diffs
> -----
> 
>   intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
>   intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityImportStream.java 73994b9 
>   repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
>   webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
>   webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 
> 
> 
> Diff: https://reviews.apache.org/r/57649/diff/1/
> 
> 
> Testing
> -------
> 
> Test data:
> - QuickStart_v1: 3 databases.
> - A *hive_db* with 922 tables.
> - Stocks *hive_db* with 1 database, table, process and 5 columns.
> - A *hive_db* with 522K entities.
> 
> The changes impact all the flows in the Export & Import APIs.
> Unit testing: Manual.
> Integration testing: Manual.
> Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.
> 
> 
> Thanks,
> 
> Ashutosh Mestry
> 
>


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Ashutosh Mestry <am...@hortonworks.com>.

> On March 16, 2017, 8:07 p.m., Madhan Neethiraj wrote:
> > webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
> > Lines 440 (patched)
> > <https://reviews.apache.org/r/57649/diff/2/?file=1666330#file1666330line467>
> >
> >     are both, add() and addUnique() needed? Given this is 'UniqueList', shouldn't add() enforce uniqueness?

Removed addUnique. add now checks for uniqueness.


- Ashutosh


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/#review169196
-----------------------------------------------------------


On March 16, 2017, 6:28 p.m., Ashutosh Mestry wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57649/
> -----------------------------------------------------------
> 
> (Updated March 16, 2017, 6:28 p.m.)
> 
> 
> Review request for atlas and Madhan Neethiraj.
> 
> 
> Bugs: ATLAS-1503
>     https://issues.apache.org/jira/browse/ATLAS-1503
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> **Background**
> ==============
> Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.
> 
> **Solution**
> ============
> The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)
> 
> This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.
> 
> **Implementation Details**
> ==========================
> *Export API*
> - Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
> - _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
> - _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.
> 
> *Import API*
> - _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
> - _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
> - _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
> - _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.
> 
> 
> **Functional Areas Impacted**
> =============================
> *Export*
> - Full
> - Connected
> - HDFS path-based import.
> 
> *Import*
> - Regular flow.
> 
> **Examples**
> ============
> Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.
> 
> Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.
> 
> **Metrics**
> ===========
> 
> Date | File Size | No. of Entities | Export   |Import   |
>      |           |                 | Duration |Duration |
> -----|-----------|-----------------|----------|---------|
> 3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
> 3/08 |      7 KB |               3 |    5 secs|    7 sec|
> --------------------------------------------------------|
>                Improvement                              | 
> --------------------------------------------------------|
> 3/14 |     38 MB |          202930 |   32 mins| 1:40 hrs| 
> 3/14 |      5 KB |               3 |    5 secs|    7 sec|
> 
> 
> **Summary**
> ===========
> With these changes the file size reduction is: ~65%.
> 
> 
> Diffs
> -----
> 
>   intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
>   intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
>   repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
>   webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
>   webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 
> 
> 
> Diff: https://reviews.apache.org/r/57649/diff/2/
> 
> 
> Testing
> -------
> 
> Test data:
> - QuickStart_v1: 3 databases.
> - A *hive_db* with 922 tables.
> - Stocks *hive_db* with 1 database, table, process and 5 columns.
> - A *hive_db* with 522K entities.
> 
> The changes impact all the flows in the Export & Import APIs.
> Unit testing: Manual.
> Integration testing: Manual.
> Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.
> 
> 
> Thanks,
> 
> Ashutosh Mestry
> 
>


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Madhan Neethiraj <ma...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/#review169196
-----------------------------------------------------------




repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java
Line 59 (original), 65 (patched)
<https://reviews.apache.org/r/57649/#comment241540>

    "new AtlasEntityWithExtInfo(iterator.next())" ==> "new AtlasEntityWithExtInfo(iterator.next(), entitiesWithExtInfo)"



repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java
Lines 35 (patched)
<https://reviews.apache.org/r/57649/#comment241541>

    "entityStream" would be null, if constructor at #23 was used.



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Lines 440 (patched)
<https://reviews.apache.org/r/57649/#comment241542>

    are both, add() and addUnique() needed? Given this is 'UniqueList', shouldn't add() enforce uniqueness?



webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java
Lines 454 (patched)
<https://reviews.apache.org/r/57649/#comment241543>

    addAll(UniqueList<T> e) ==> this might endup with duplicate entries in 'list'. If this method is needed, consider iterating the list and call add(element) for each element.


- Madhan Neethiraj


On March 16, 2017, 6:28 p.m., Ashutosh Mestry wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57649/
> -----------------------------------------------------------
> 
> (Updated March 16, 2017, 6:28 p.m.)
> 
> 
> Review request for atlas and Madhan Neethiraj.
> 
> 
> Bugs: ATLAS-1503
>     https://issues.apache.org/jira/browse/ATLAS-1503
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> **Background**
> ==============
> Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.
> 
> **Solution**
> ============
> The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)
> 
> This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.
> 
> **Implementation Details**
> ==========================
> *Export API*
> - Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
> - _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
> - _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.
> 
> *Import API*
> - _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
> - _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
> - _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
> - _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.
> 
> 
> **Functional Areas Impacted**
> =============================
> *Export*
> - Full
> - Connected
> - HDFS path-based import.
> 
> *Import*
> - Regular flow.
> 
> **Examples**
> ============
> Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.
> 
> Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.
> 
> **Metrics**
> ===========
> 
> Date | File Size | No. of Entities | Export   |Import   |
>      |           |                 | Duration |Duration |
> -----|-----------|-----------------|----------|---------|
> 3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
> 3/08 |      7 KB |               3 |    5 secs|    7 sec|
> --------------------------------------------------------|
>                Improvement                              | 
> --------------------------------------------------------|
> 3/14 |     38 MB |          202930 |   32 mins| 1:40 hrs| 
> 3/14 |      5 KB |               3 |    5 secs|    7 sec|
> 
> 
> **Summary**
> ===========
> With these changes the file size reduction is: ~65%.
> 
> 
> Diffs
> -----
> 
>   intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
>   intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
>   repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
>   webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
>   webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 
> 
> 
> Diff: https://reviews.apache.org/r/57649/diff/2/
> 
> 
> Testing
> -------
> 
> Test data:
> - QuickStart_v1: 3 databases.
> - A *hive_db* with 922 tables.
> - Stocks *hive_db* with 1 database, table, process and 5 columns.
> - A *hive_db* with 522K entities.
> 
> The changes impact all the flows in the Export & Import APIs.
> Unit testing: Manual.
> Integration testing: Manual.
> Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.
> 
> 
> Thanks,
> 
> Ashutosh Mestry
> 
>


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Apoorv Naik <na...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/#review169383
-----------------------------------------------------------



Jenkins is going to check this review request...

- Apoorv Naik


On March 17, 2017, 5:09 a.m., Ashutosh Mestry wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57649/
> -----------------------------------------------------------
> 
> (Updated March 17, 2017, 5:09 a.m.)
> 
> 
> Review request for atlas and Madhan Neethiraj.
> 
> 
> Bugs: ATLAS-1665
>     https://issues.apache.org/jira/browse/ATLAS-1665
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> **Background**
> ==============
> Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.
> 
> **Solution**
> ============
> The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)
> 
> This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.
> 
> **Implementation Details**
> ==========================
> *Export API*
> - Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
> - _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
> - _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.
> 
> *Import API*
> - _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
> - _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
> - _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
> - _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.
> 
> 
> **Functional Areas Impacted**
> =============================
> *Export*
> - Full
> - Connected
> - HDFS path-based import.
> 
> *Import*
> - Regular flow.
> 
> **Examples**
> ============
> Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.
> 
> Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.
> 
> **Metrics**
> ===========
> 
> Date | File Size | No. of Entities | Export   |Import   |
>      |           |                 | Duration |Duration |
> -----|-----------|-----------------|----------|---------|
> 3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
> 3/08 |      7 KB |               3 |    5 secs|    7 sec|
> --------------------------------------------------------|
>                Improvement                              | 
> --------------------------------------------------------|
> 3/14 |     38 MB |          202930 |   20 mins| 1:10 hrs| 
> 3/14 |      5 KB |               3 |    5 secs|    7 sec|
> 
> 
> **Summary**
> ===========
> With these changes the file size reduction is: ~65%.
> 
> 
> Diffs
> -----
> 
>   intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
>   repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
>   webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java e123ff7 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 37d9eb5 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 
> 
> 
> Diff: https://reviews.apache.org/r/57649/diff/6/
> 
> 
> Testing
> -------
> 
> Test data:
> - QuickStart_v1: 3 databases.
> - A *hive_db* with 922 tables.
> - Stocks *hive_db* with 1 database, table, process and 5 columns.
> - A *hive_db* with 522K entities.
> 
> The changes impact all the flows in the Export & Import APIs.
> Unit testing: Manual.
> Integration testing: Manual.
> Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.
> 
> 
> File Attachments
> ----------------
> 
> Patch on 2.6-maint
>   https://reviews.apache.org/media/uploaded/files/2017/03/17/5fc7a466-9bac-4282-a9fd-659d0528b443__export-size-optimized.2.6-maint.2.patch
> 
> 
> Thanks,
> 
> Ashutosh Mestry
> 
>


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Madhan Neethiraj <ma...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/#review169370
-----------------------------------------------------------


Ship it!




Ship It!

- Madhan Neethiraj


On March 17, 2017, 5:09 a.m., Ashutosh Mestry wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57649/
> -----------------------------------------------------------
> 
> (Updated March 17, 2017, 5:09 a.m.)
> 
> 
> Review request for atlas and Madhan Neethiraj.
> 
> 
> Bugs: ATLAS-1665
>     https://issues.apache.org/jira/browse/ATLAS-1665
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> **Background**
> ==============
> Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.
> 
> **Solution**
> ============
> The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)
> 
> This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.
> 
> **Implementation Details**
> ==========================
> *Export API*
> - Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
> - _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
> - _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.
> 
> *Import API*
> - _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
> - _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
> - _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
> - _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.
> 
> 
> **Functional Areas Impacted**
> =============================
> *Export*
> - Full
> - Connected
> - HDFS path-based import.
> 
> *Import*
> - Regular flow.
> 
> **Examples**
> ============
> Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.
> 
> Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.
> 
> **Metrics**
> ===========
> 
> Date | File Size | No. of Entities | Export   |Import   |
>      |           |                 | Duration |Duration |
> -----|-----------|-----------------|----------|---------|
> 3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
> 3/08 |      7 KB |               3 |    5 secs|    7 sec|
> --------------------------------------------------------|
>                Improvement                              | 
> --------------------------------------------------------|
> 3/14 |     38 MB |          202930 |   20 mins| 1:10 hrs| 
> 3/14 |      5 KB |               3 |    5 secs|    7 sec|
> 
> 
> **Summary**
> ===========
> With these changes the file size reduction is: ~65%.
> 
> 
> Diffs
> -----
> 
>   intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
>   repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
>   webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java e123ff7 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 37d9eb5 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 
> 
> 
> Diff: https://reviews.apache.org/r/57649/diff/6/
> 
> 
> Testing
> -------
> 
> Test data:
> - QuickStart_v1: 3 databases.
> - A *hive_db* with 922 tables.
> - Stocks *hive_db* with 1 database, table, process and 5 columns.
> - A *hive_db* with 522K entities.
> 
> The changes impact all the flows in the Export & Import APIs.
> Unit testing: Manual.
> Integration testing: Manual.
> Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.
> 
> 
> File Attachments
> ----------------
> 
> Patch on 2.6-maint
>   https://reviews.apache.org/media/uploaded/files/2017/03/17/5fc7a466-9bac-4282-a9fd-659d0528b443__export-size-optimized.2.6-maint.2.patch
> 
> 
> Thanks,
> 
> Ashutosh Mestry
> 
>


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Apoorv Naik <na...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/#review169384
-----------------------------------------------------------



Uh oh, Jenkins tried to build the changes and failed. See [atlas-UTs #1](http://osboxes:8080/job/atlas-UTs/1/).

- Apoorv Naik


On March 17, 2017, 5:09 a.m., Ashutosh Mestry wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57649/
> -----------------------------------------------------------
> 
> (Updated March 17, 2017, 5:09 a.m.)
> 
> 
> Review request for atlas and Madhan Neethiraj.
> 
> 
> Bugs: ATLAS-1665
>     https://issues.apache.org/jira/browse/ATLAS-1665
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> **Background**
> ==============
> Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.
> 
> **Solution**
> ============
> The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)
> 
> This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.
> 
> **Implementation Details**
> ==========================
> *Export API*
> - Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
> - _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
> - _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.
> 
> *Import API*
> - _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
> - _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
> - _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
> - _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.
> 
> 
> **Functional Areas Impacted**
> =============================
> *Export*
> - Full
> - Connected
> - HDFS path-based import.
> 
> *Import*
> - Regular flow.
> 
> **Examples**
> ============
> Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.
> 
> Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.
> 
> **Metrics**
> ===========
> 
> Date | File Size | No. of Entities | Export   |Import   |
>      |           |                 | Duration |Duration |
> -----|-----------|-----------------|----------|---------|
> 3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
> 3/08 |      7 KB |               3 |    5 secs|    7 sec|
> --------------------------------------------------------|
>                Improvement                              | 
> --------------------------------------------------------|
> 3/14 |     38 MB |          202930 |   20 mins| 1:10 hrs| 
> 3/14 |      5 KB |               3 |    5 secs|    7 sec|
> 
> 
> **Summary**
> ===========
> With these changes the file size reduction is: ~65%.
> 
> 
> Diffs
> -----
> 
>   intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
>   repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
>   repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
>   webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java e123ff7 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 37d9eb5 
>   webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 
> 
> 
> Diff: https://reviews.apache.org/r/57649/diff/6/
> 
> 
> Testing
> -------
> 
> Test data:
> - QuickStart_v1: 3 databases.
> - A *hive_db* with 922 tables.
> - Stocks *hive_db* with 1 database, table, process and 5 columns.
> - A *hive_db* with 522K entities.
> 
> The changes impact all the flows in the Export & Import APIs.
> Unit testing: Manual.
> Integration testing: Manual.
> Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.
> 
> 
> File Attachments
> ----------------
> 
> Patch on 2.6-maint
>   https://reviews.apache.org/media/uploaded/files/2017/03/17/5fc7a466-9bac-4282-a9fd-659d0528b443__export-size-optimized.2.6-maint.2.patch
> 
> 
> Thanks,
> 
> Ashutosh Mestry
> 
>


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Ashutosh Mestry <am...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/
-----------------------------------------------------------

(Updated March 17, 2017, 5:09 a.m.)


Review request for atlas and Madhan Neethiraj.


Changes
-------

Export performance improvement by ~35%.


Bugs: ATLAS-1503
    https://issues.apache.org/jira/browse/ATLAS-1503


Repository: atlas


Description
-------

**Background**
==============
Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.

**Solution**
============
The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)

This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.

**Implementation Details**
==========================
*Export API*
- Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
- _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
- _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.

*Import API*
- _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
- _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
- _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
- _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.


**Functional Areas Impacted**
=============================
*Export*
- Full
- Connected
- HDFS path-based import.

*Import*
- Regular flow.

**Examples**
============
Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.

Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.

**Metrics**
===========

Date | File Size | No. of Entities | Export   |Import   |
     |           |                 | Duration |Duration |
-----|-----------|-----------------|----------|---------|
3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
3/08 |      7 KB |               3 |    5 secs|    7 sec|
--------------------------------------------------------|
               Improvement                              | 
--------------------------------------------------------|
3/14 |     38 MB |          202930 |   20 mins| 1:40 hrs| 
3/14 |      5 KB |               3 |    5 secs|    7 sec|


**Summary**
===========
With these changes the file size reduction is: ~65%.


Diffs (updated)
-----

  intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
  intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
  repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
  webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
  webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 


Diff: https://reviews.apache.org/r/57649/diff/4/

Changes: https://reviews.apache.org/r/57649/diff/3-4/


Testing
-------

Test data:
- QuickStart_v1: 3 databases.
- A *hive_db* with 922 tables.
- Stocks *hive_db* with 1 database, table, process and 5 columns.
- A *hive_db* with 522K entities.

The changes impact all the flows in the Export & Import APIs.
Unit testing: Manual.
Integration testing: Manual.
Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.


File Attachments
----------------

Patch on 2.6-maint
  https://reviews.apache.org/media/uploaded/files/2017/03/17/25235e5a-180b-464e-96d6-d7e45a157152__export-size-optimized.2.6-maint.patch


Thanks,

Ashutosh Mestry


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Ashutosh Mestry <am...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/
-----------------------------------------------------------

(Updated March 16, 2017, 10:37 p.m.)


Review request for atlas and Madhan Neethiraj.


Changes
-------

Contents:
- Fixed export progress reporting.
- Addressed review comments.


Bugs: ATLAS-1503
    https://issues.apache.org/jira/browse/ATLAS-1503


Repository: atlas


Description
-------

**Background**
==============
Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.

**Solution**
============
The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)

This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.

**Implementation Details**
==========================
*Export API*
- Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
- _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
- _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.

*Import API*
- _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
- _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
- _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
- _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.


**Functional Areas Impacted**
=============================
*Export*
- Full
- Connected
- HDFS path-based import.

*Import*
- Regular flow.

**Examples**
============
Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.

Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.

**Metrics**
===========

Date | File Size | No. of Entities | Export   |Import   |
     |           |                 | Duration |Duration |
-----|-----------|-----------------|----------|---------|
3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
3/08 |      7 KB |               3 |    5 secs|    7 sec|
--------------------------------------------------------|
               Improvement                              | 
--------------------------------------------------------|
3/14 |     38 MB |          202930 |   32 mins| 1:40 hrs| 
3/14 |      5 KB |               3 |    5 secs|    7 sec|


**Summary**
===========
With these changes the file size reduction is: ~65%.


Diffs (updated)
-----

  intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
  intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
  repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
  webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
  webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 


Diff: https://reviews.apache.org/r/57649/diff/3/

Changes: https://reviews.apache.org/r/57649/diff/2-3/


Testing
-------

Test data:
- QuickStart_v1: 3 databases.
- A *hive_db* with 922 tables.
- Stocks *hive_db* with 1 database, table, process and 5 columns.
- A *hive_db* with 522K entities.

The changes impact all the flows in the Export & Import APIs.
Unit testing: Manual.
Integration testing: Manual.
Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.


Thanks,

Ashutosh Mestry


Re: Review Request 57649: Export API: ZIP File Size Optimization

Posted by Ashutosh Mestry <am...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57649/
-----------------------------------------------------------

(Updated March 16, 2017, 6:28 p.m.)


Review request for atlas and Madhan Neethiraj.


Changes
-------

Addressed review comments.


Bugs: ATLAS-1503
    https://issues.apache.org/jira/browse/ATLAS-1503


Repository: atlas


Description (updated)
-------

**Background**
==============
Existing implementation of Export API w.r.t ZIP file genration adds 1 *.json* file per entitiy. This makes ZIP file creation inefficient. The ZIP files are 75% larger in size than what could be possible with fewer *.json* file entries.

**Solution**
============
The implementation uses the new v2 API *AtlasEntityWithExtInfo* representation instead of *AtlasEntity*. This format combines an entity with related entities as one. E.g. *hive_table* will contain all the *hive_columns* that it is made up of. (See example section below.)

This results in significant reduction of generated *JSON* files. This impacts reduction in generated *ZIP* file.

**Implementation Details**
==========================
*Export API*
- Modified *Gremlin* used to fetch connected entities to return *guid* with *boolean* to indicate if the entity is process or not.
- _ExportService_ Modified implementation to fetch *AtlasEntityWithExtInfo* instead of *AtlasEntity*. Modified book keeping to save *process* (lineage) entities after all non-process entities are saved.
- _ZipSink_ Minor modification to serialize  *AtlasEntityWithExtInfo*.

*Import API*
- _ZipSource_ Modified to source *AtlasEntityWithExtInfo*.
- _EntityImportStream_ Modified to source *AtlasEntityWithExtInfo*.
- _AtlasEntityStreamForImport.getGuid_ Modified  to source requested entities first from stored *AtlasEntityWithExtInfo* object. Request from stream only if not found.
- _AtlasEntityStoreV1.bulkImport_ Minor modification to use the new changes to stream.


**Functional Areas Impacted**
=============================
*Export*
- Full
- Connected
- HDFS path-based import.

*Import*
- Regular flow.

**Examples**
============
Case *hive_db*: Within the GraphDB the database has inward edges from objects that refer to it. Tables in this case. So *AtlasEntityWithExtInfo* for database will not have any referred entities.

Case of *hive_table*: Within the GraphDB the table has outward edges pointing to the columns it is made up of. It also has edges pointing to database and storage descriptor. Hence, the *AtlasEntityWithExtInfo* for table will have all full representation of all the columns and reference to database and storage descriptor.

**Metrics**
===========

Date | File Size | No. of Entities | Export   |Import   |
     |           |                 | Duration |Duration |
-----|-----------|-----------------|----------|---------|
3/02 |    180 MB |          202930 |   22 mins| 1:38 hrs|
3/08 |      7 KB |               3 |    5 secs|    7 sec|
--------------------------------------------------------|
               Improvement                              | 
--------------------------------------------------------|
3/14 |     38 MB |          202930 |   32 mins| 1:40 hrs| 
3/14 |      5 KB |               3 |    5 secs|    7 sec|


**Summary**
===========
With these changes the file size reduction is: ~65%.


Diffs (updated)
-----

  intg/src/main/java/org/apache/atlas/model/impexp/AtlasExportResult.java e6a967e 
  intg/src/main/java/org/apache/atlas/model/instance/AtlasEntity.java 4e3895d 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityGraphDiscoveryV1.java 6c88510 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStoreV1.java cce3fca 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStream.java 5d9a7d4 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/AtlasEntityStreamForImport.java 8cb36ac 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/EntityStream.java 4c43921 
  repository/src/main/java/org/apache/atlas/repository/store/graph/v1/InMemoryMapEntityStream.java 241f6d0 
  repository/src/main/java/org/apache/atlas/util/AtlasGremlin2QueryProvider.java 4743b73 
  webapp/src/main/java/org/apache/atlas/web/resources/AdminResource.java 31a4cf9 
  webapp/src/main/java/org/apache/atlas/web/resources/ExportService.java c1891e0 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSink.java 2e4cb01 
  webapp/src/main/java/org/apache/atlas/web/resources/ZipSource.java a69f7fa 


Diff: https://reviews.apache.org/r/57649/diff/2/

Changes: https://reviews.apache.org/r/57649/diff/1-2/


Testing
-------

Test data:
- QuickStart_v1: 3 databases.
- A *hive_db* with 922 tables.
- Stocks *hive_db* with 1 database, table, process and 5 columns.
- A *hive_db* with 522K entities.

The changes impact all the flows in the Export & Import APIs.
Unit testing: Manual.
Integration testing: Manual.
Accuracy testing: Manual. Verified using Export -> Import -> Export -> file compare.


Thanks,

Ashutosh Mestry