GitHub actions for testing

[darabos]

For #8. It's hard to test this locally. I'm using https://github.com/nektos/act but I'm getting weird errors and caching doesn't work, so each attempt takes ages. Will this PR trigger a run, I wonder? If not, I may merge this and try to see if I can trigger it that way.

[darabos]

Will this PR trigger a run, I wonder?

Looks that way! I can see this PR on the Actions page!

[darabos]

It dies with the same weird error:

[error] /github/workspace/app/Global.scala:56: object io is not a member of package org.apache.commons
[error]           org.apache.commons.io.FileUtils.writeStringToFile(
[error]                              ^
[error] /github/workspace/app/com/lynxanalytics/biggraph/BigGraphEnvironment.scala:13: type SparkSession is not a member of package org.apache.spark.sql
[error]   def createSparkSession: spark.sql.SparkSession

But it goes quite fast, just 14 minutes. Looks like maybe the Spark import does not work out.

[xandrew]

Thanks, promising! Shall I review or wait until you got to the bottom of the compile issue?

[darabos]

Thanks, promising! Shall I review or wait until you got to the bottom of the compile issue?

Thanks! Yeah, don't review it yet.

[darabos]

Now it's actually running our tests! Most of them even pass. I've seen a failure that I cannot reproduce locally though:

[info] - Derive graph attribute *** FAILED ***
[info]   java.nio.charset.MalformedInputException: Input length = 1
[info]   at java.nio.charset.CoderResult.throwException(CoderResult.java:281)
[info]   at sun.nio.cs.StreamDecoder.implRead(StreamDecoder.java:339)
[info]   at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:178)
[info]   at java.io.InputStreamReader.read(InputStreamReader.java:184)
[info]   at java.io.BufferedReader.fill(BufferedReader.java:161)
[info]   at java.io.BufferedReader.readLine(BufferedReader.java:324)
[info]   at java.io.BufferedReader.readLine(BufferedReader.java:389)
[info]   at scala.io.BufferedSource$BufferedLineIterator.hasNext(BufferedSource.scala:72)
[info]   at scala.collection.Iterator$class.foreach(Iterator.scala:893)
[info]   at scala.collection.AbstractIterator.foreach(Iterator.scala:1336)
[info]   ...

[darabos]

[error] Error: Total 690, Failed 13, Errors 2, Passed 675, Ignored 4
[error] Failed tests:
[error] 	com.lynxanalytics.biggraph.frontend_operations.ComputeInPythonTest
[error] 	com.lynxanalytics.biggraph.graph_operations.GCNTest
[error] 	com.lynxanalytics.biggraph.frontend_operations.DeriveGraphAttributeTest
[error] 	com.lynxanalytics.biggraph.graph_operations.DeriveScalaTest
[error] 	com.lynxanalytics.biggraph.frontend_operations.EmbedVerticesTest
[error] 	com.lynxanalytics.biggraph.frontend_operations.SQLTest
[error] 	com.lynxanalytics.biggraph.frontend_operations.CreateGraphInPythonTest
[error] Error during tests:
[error] 	com.lynxanalytics.biggraph.frontend_operations.ImportNeo4jTest
[error] 	com.lynxanalytics.biggraph.graph_util.Neo4jUtilTest
[error] (test:testOnly) sbt.TestsFailedException: Tests unsuccessful
[error] Total time: 1511 s, completed Jul 10, 2020 1:11:51 PM

[darabos]

The backend test passes now!

I'm not sure what the frontend test is doing. If I click on it to see the logs, nothing happens. Surely it will time out or something eventually...

[darabos]

Hallelujah! 🎉

I'm sure we can improve them but it's a start!

• They are a bit slow. I'm not sure the caching is working right. It's a shame that the Docker image gets rebuilt all the time.
• They are not very reliable. When SBT downloads the dependencies (e.g. SBT itself), I've seen multiple cases of timeouts. If the cache worked that would help with this too.

But I think we can merge it for now!

[xandrew]

Can't say I thoroughly understood all details, but looks good in general! Cool that we have a working CI now! :)

[xandrew]

Pytorch is our only dependency? Maybe worth renaming this file? I mean, someone might innocently add something at the end and get surprised by it not working, due to the if at the start. Or maybe reverse the if. and put the torch related pip installs inside? Btw, why the if? Doesn't pip handle this anyways?

[darabos]

Pytorch is our only dependency? Maybe worth renaming this file? I mean, someone might innocently add something at the end and get surprised by it not working, due to the if at the start. Or maybe reverse the if. and put the torch related pip installs inside? Btw, why the if? Doesn't pip handle this anyways?

Without the if it takes 40 seconds to do nothing. We could probably use Pipenv or some magic to handle this better, but I'm just happy it works at all. 😅

I've reversed the if, thanks for the suggestion!