This is an automated email from the ASF dual-hosted git repository.
jiwq pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/submarine.git
The following commit(s) were added to refs/heads/master by this push:
new 3968cf4 SUBMARINE-542. [SDK] get_log error when experiment is not
started
3968cf4 is described below
commit 3968cf4375bcc2098d926bb55acac17543b09b44
Author: pingsutw <[email protected]>
AuthorDate: Wed Jun 24 15:08:10 2020 +0800
SUBMARINE-542. [SDK] get_log error when experiment is not started
### What is this PR for?
Fix get_log error when the experiment is not started
```bash
IndexError Traceback (most recent call last)
<ipython-input-13-8bf55eff97d1> in <module>
----> 1
submarine_client.get_log(id)~/opt/anaconda3/envs/python3-7/lib/python3.7/site-packages/submarine/experiment/api/experiment_client.py
in get_log(self, id, master)
124
125 if master is True:
--> 126 log_contents = [log_contents[0]]
127
128 for log_content in log_contents:IndexError: list index out
of range
```
### What type of PR is it?
[Bug Fix]
### Todos
* [ ] - Task
### What is the Jira issue?
https://issues.apache.org/jira/browse/SUBMARINE-542
### How should this be tested?
https://travis-ci.org/github/pingsutw/hadoop-submarine/builds/701089803
### Screenshots (if appropriate)
### Questions:
* Does the licenses files need update? No
* Is there breaking changes for older versions? No
* Does this needs documentation? No
Author: pingsutw <[email protected]>
Author: HUAN-PING SU <[email protected]>
Closes #323 from pingsutw/SUBMARINE-542 and squashes the following commits:
4620ea3 [pingsutw] Fix checkstyle
92ae7df [HUAN-PING SU] Update experiment_client.py
ee63190 [pingsutw] Fix code style
ce458ad [pingsutw] SUBMARINE-542. [SDK] get_log error when experiment is
pulling docker images
---
.../submarine/experiment/api/experiment_client.py | 21 ++++++++++++---------
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git
a/submarine-sdk/pysubmarine/submarine/experiment/api/experiment_client.py
b/submarine-sdk/pysubmarine/submarine/experiment/api/experiment_client.py
index e8272a9..00adc1d 100644
--- a/submarine-sdk/pysubmarine/submarine/experiment/api/experiment_client.py
+++ b/submarine-sdk/pysubmarine/submarine/experiment/api/experiment_client.py
@@ -65,7 +65,10 @@ class ExperimentClient:
def _log_pod(self, id, index):
response = self.experiment_api.get_log(id)
- log_content = response.result['logContent'][0]
+ log_contents = response.result['logContent']
+ if len(log_contents) == 0:
+ return index
+ log_content = log_contents[0]
for i, log in enumerate(log_content['podLog']):
if i < index:
continue
@@ -110,19 +113,19 @@ class ExperimentClient:
response = self.experiment_api.delete_experiment(id)
return response.result
- def get_log(self, id, master=True):
+ def get_log(self, id, onlyMaster=False):
"""
- Get training logs of the experiment.
- By default only get the logs of Pod that has labels 'job-role: master'.
- :param master: By default get pod with label 'job-role: master' pod if
True.
- If need to get more Pod Logs, set False.
- :param id: Experiment ID
- :return: str: experiment logs
+ Get training logs of all pod of the experiment.
+ By default get all the logs of Pod
+ :param id: experiment id
+ :param onlyMaster: By default include pod log of "master" which might
be
+ Tensorflow PS/Chief or PyTorch master
+ :return: str: pods logs
"""
response = self.experiment_api.get_log(id)
log_contents = response.result['logContent']
- if master is True:
+ if onlyMaster is True and len(log_contents) != 0:
log_contents = [log_contents[0]]
for log_content in log_contents:
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
