| *** bpokorny_ has joined #openstack-searchlight | 00:00 | |
| *** bpokorny has quit IRC | 00:04 | |
| *** lei-zh- has quit IRC | 00:04 | |
| *** lakshmiS has joined #openstack-searchlight | 00:11 | |
| openstackgerrit | Travis Tripp proposed openstack/searchlight: Add Devstack install support for searchlightclient https://review.openstack.org/268358 | 00:20 |
|---|---|---|
| openstackgerrit | Travis Tripp proposed openstack/searchlight: Add Devstack install support for searchlightclient https://review.openstack.org/268358 | 00:21 |
| openstackgerrit | Li Yingjun proposed openstack/python-searchlightclient: Add facet list support https://review.openstack.org/255027 | 00:37 |
| openstackgerrit | Li Yingjun proposed openstack/python-searchlightclient: Added search resource client and cli https://review.openstack.org/260899 | 00:37 |
| openstackgerrit | Li Yingjun proposed openstack/python-searchlightclient: Add resource type list CLI https://review.openstack.org/249076 | 00:37 |
| openstackgerrit | Merged openstack/python-searchlightclient: Add resource type list CLI https://review.openstack.org/249076 | 00:50 |
| openstackgerrit | Merged openstack/python-searchlightclient: Add facet list support https://review.openstack.org/255027 | 00:52 |
| *** bpokorny_ has quit IRC | 00:55 | |
| *** bpokorny has joined #openstack-searchlight | 00:55 | |
| *** bpokorny has quit IRC | 01:01 | |
| *** yingjun has joined #openstack-searchlight | 01:07 | |
| *** lakshmiS has quit IRC | 01:10 | |
| *** TravT has quit IRC | 01:51 | |
| *** TravT has joined #openstack-searchlight | 01:52 | |
| *** TravT has quit IRC | 01:52 | |
| *** TravT has joined #openstack-searchlight | 01:52 | |
| *** TravT has quit IRC | 01:57 | |
| *** bpokorny has joined #openstack-searchlight | 02:48 | |
| *** TravT has joined #openstack-searchlight | 02:58 | |
| *** TravT has quit IRC | 03:06 | |
| *** bpokorny_ has joined #openstack-searchlight | 03:17 | |
| *** bpokorny has quit IRC | 03:21 | |
| *** bpokorny_ has quit IRC | 03:22 | |
| *** TravT has joined #openstack-searchlight | 03:31 | |
| *** TravT has quit IRC | 03:36 | |
| *** TravT has joined #openstack-searchlight | 03:40 | |
| *** TravT has quit IRC | 03:49 | |
| *** TravT has joined #openstack-searchlight | 03:54 | |
| *** TravT has quit IRC | 04:08 | |
| *** TravT has joined #openstack-searchlight | 04:13 | |
| *** TravT has quit IRC | 04:17 | |
| *** TravT has joined #openstack-searchlight | 04:19 | |
| *** TravT has quit IRC | 04:25 | |
| *** TravT has joined #openstack-searchlight | 04:28 | |
| *** TravT has quit IRC | 04:37 | |
| *** bpokorny has joined #openstack-searchlight | 04:38 | |
| *** TravT has joined #openstack-searchlight | 04:42 | |
| *** TravT has quit IRC | 04:47 | |
| *** TravT has joined #openstack-searchlight | 04:52 | |
| *** TravT has quit IRC | 04:57 | |
| *** TravT has joined #openstack-searchlight | 04:57 | |
| *** TravT has quit IRC | 05:02 | |
| *** david-lyle has quit IRC | 05:02 | |
| *** david-lyle has joined #openstack-searchlight | 05:03 | |
| *** TravT has joined #openstack-searchlight | 05:07 | |
| *** TravT has quit IRC | 05:16 | |
| *** TravT has joined #openstack-searchlight | 05:21 | |
| *** TravT_ has joined #openstack-searchlight | 05:35 | |
| *** TravT has quit IRC | 05:35 | |
| *** TravT_ has quit IRC | 05:44 | |
| *** TravT has joined #openstack-searchlight | 05:49 | |
| *** TravT has quit IRC | 05:54 | |
| *** TravT has joined #openstack-searchlight | 06:08 | |
| *** bpokorny has quit IRC | 06:12 | |
| *** TravT has quit IRC | 06:12 | |
| *** TravT has joined #openstack-searchlight | 06:17 | |
| *** TravT has quit IRC | 06:22 | |
| *** TravT has joined #openstack-searchlight | 06:27 | |
| *** TravT has quit IRC | 06:31 | |
| *** TravT has joined #openstack-searchlight | 06:32 | |
| *** TravT has quit IRC | 06:37 | |
| *** TravT has joined #openstack-searchlight | 06:41 | |
| *** TravT has quit IRC | 06:46 | |
| *** TravT has joined #openstack-searchlight | 06:51 | |
| *** TravT has quit IRC | 06:55 | |
| openstackgerrit | Andreas Jaeger proposed openstack/python-searchlightclient: Remove argparse from requirements https://review.openstack.org/270390 | 06:56 |
| *** TravT has joined #openstack-searchlight | 07:00 | |
| *** lakshmiS has joined #openstack-searchlight | 07:06 | |
| *** lakshmiS has quit IRC | 07:12 | |
| *** TravT has quit IRC | 07:19 | |
| *** TravT has joined #openstack-searchlight | 07:23 | |
| *** TravT has quit IRC | 07:28 | |
| *** TravT has joined #openstack-searchlight | 07:33 | |
| *** TravT has quit IRC | 07:37 | |
| *** TravT has joined #openstack-searchlight | 08:25 | |
| *** TravT has quit IRC | 08:39 | |
| *** TravT has joined #openstack-searchlight | 08:44 | |
| *** TravT has quit IRC | 08:48 | |
| *** exploreshaifali has joined #openstack-searchlight | 09:31 | |
| *** yingjun has quit IRC | 09:33 | |
| *** TravT has joined #openstack-searchlight | 09:36 | |
| *** TravT has quit IRC | 09:41 | |
| *** TravT has joined #openstack-searchlight | 09:45 | |
| *** TravT has quit IRC | 09:50 | |
| *** TravT has joined #openstack-searchlight | 09:55 | |
| *** TravT has quit IRC | 10:00 | |
| *** openstackgerrit has quit IRC | 10:02 | |
| *** openstackgerrit has joined #openstack-searchlight | 10:03 | |
| *** TravT has joined #openstack-searchlight | 10:05 | |
| *** TravT has quit IRC | 10:10 | |
| *** TravT has joined #openstack-searchlight | 10:29 | |
| *** TravT has quit IRC | 10:38 | |
| *** TravT has joined #openstack-searchlight | 10:58 | |
| *** TravT has quit IRC | 11:02 | |
| *** TravT has joined #openstack-searchlight | 11:07 | |
| *** TravT has quit IRC | 11:21 | |
| *** TravT has joined #openstack-searchlight | 11:25 | |
| *** TravT has quit IRC | 11:30 | |
| *** TravT has joined #openstack-searchlight | 11:31 | |
| *** TravT has quit IRC | 11:35 | |
| *** yingjun has joined #openstack-searchlight | 11:37 | |
| *** TravT has joined #openstack-searchlight | 11:40 | |
| *** TravT has quit IRC | 11:44 | |
| *** TravT has joined #openstack-searchlight | 11:45 | |
| *** TravT has quit IRC | 11:54 | |
| *** TravT has joined #openstack-searchlight | 12:18 | |
| *** TravT has quit IRC | 12:23 | |
| *** TravT has joined #openstack-searchlight | 12:28 | |
| *** TravT has quit IRC | 12:42 | |
| *** TravT has joined #openstack-searchlight | 12:47 | |
| *** TravT has quit IRC | 12:56 | |
| *** TravT has joined #openstack-searchlight | 13:00 | |
| *** TravT has quit IRC | 13:05 | |
| *** TravT has joined #openstack-searchlight | 13:10 | |
| *** TravT has quit IRC | 13:24 | |
| *** TravT has joined #openstack-searchlight | 13:25 | |
| *** TravT has quit IRC | 13:29 | |
| *** TravT has joined #openstack-searchlight | 13:34 | |
| *** TravT has quit IRC | 13:38 | |
| *** TravT has joined #openstack-searchlight | 13:44 | |
| *** TravT has quit IRC | 13:48 | |
| *** TravT has joined #openstack-searchlight | 13:53 | |
| *** exploreshaifali has quit IRC | 13:54 | |
| *** TravT has quit IRC | 13:58 | |
| *** TravT has joined #openstack-searchlight | 14:03 | |
| *** TravT has quit IRC | 14:12 | |
| *** TravT_ has joined #openstack-searchlight | 14:12 | |
| *** TravT_ has quit IRC | 14:17 | |
| *** lakshmiS has joined #openstack-searchlight | 14:30 | |
| *** TravT has joined #openstack-searchlight | 14:42 | |
| *** TravT has quit IRC | 14:51 | |
| *** TravT has joined #openstack-searchlight | 14:52 | |
| TravT | Courtesy Searchlight meeting reminder in #openstack-meeting-4: lakshmiS, nikhil_k, rosmaita, TravT, david-lyle, kragniz, sjmc7, ekarlso, abhijeetm, itisha, briancline, lei-zh, yingjun | 14:59 |
| sjmc7 | ta | 15:01 |
| rosmaita | o/ | 15:01 |
| TravT | hey rosmaita: perhaps you meant to wave o/ in the other room ;) | 15:01 |
| rosmaita | oops | 15:02 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 15:03 | |
| sjmc7 | you’re just so happy! | 15:14 |
| *** TravT_ has joined #openstack-searchlight | 15:21 | |
| *** TravT has quit IRC | 15:25 | |
| *** TravT_ is now known as TravT | 15:25 | |
| *** TravT_ has joined #openstack-searchlight | 15:34 | |
| *** TravT has quit IRC | 15:37 | |
| *** TravT_ is now known as TravT | 15:38 | |
| *** lakshmiS has quit IRC | 16:00 | |
| *** dhellmann has joined #openstack-searchlight | 16:07 | |
| openstackgerrit | Li Yingjun proposed openstack/python-searchlightclient: Remove dash for resource-type in cli https://review.openstack.org/270894 | 16:20 |
| *** yingjun has quit IRC | 16:25 | |
| *** bpokorny has joined #openstack-searchlight | 16:44 | |
| *** bpokorny has quit IRC | 16:51 | |
| *** RickA-HP has joined #openstack-searchlight | 16:51 | |
| *** bpokorny has joined #openstack-searchlight | 16:52 | |
| *** bpokorny has quit IRC | 16:53 | |
| sjmc7 | TravT: RickA-HP : rather than playing gerrit tag on https://review.openstack.org/#/c/245222/8/specs/mitaka/zero-downtime-reindexing.rst, i think i’m still a little confused around assigning aliases to resource types. if two resource types, ‘A’ and ‘B’ are both on the same logical index ‘index-1’ which has an alias ‘index’, i don’t think it’s possible to reindex and reassign the alias for type ‘A’ but not ‘B' | 17:00 |
| sjmc7 | which is what i mean when i say that i don’t think it makes sense to have an alias per resource type | 17:00 |
| TravT | i think we need a diagram or whiteboard or something | 17:01 |
| sjmc7 | i didn’t state the example terribly well, re-reading it | 17:01 |
| sjmc7 | ok, so ‘A’ and ‘B’ has their aliases, a-index and b-index | 17:01 |
| sjmc7 | i decide to reindex A (or both) | 17:01 |
| sjmc7 | and at some popint, a-index and b-index end up oiinting at different logical indices | 17:02 |
| sjmc7 | now there are two copies of ‘A’ and ‘B’s data | 17:02 |
| sjmc7 | it’s not possible to run a query against a-index and b-index and differentiate which copies you’ll query | 17:02 |
| RickA-HP | Too make sure I understand this, I'm going to rename thing. Resource Type (RT) A has an alias "aliasA" which points to index "index1". RT B has an alias "aliasB" which also points to index "index1". Is this correct so far? | 17:04 |
| sjmc7 | right | 17:04 |
| RickA-HP | Now RT A decides to re-index. We create a new index "index2" We update "aliasA" to point to both "index1" and "index2". | 17:05 |
| sjmc7 | umm.. no | 17:05 |
| sjmc7 | we flip aliasA to point at index2 once it’s done | 17:05 |
| RickA-HP | I haven't got that far yet :) | 17:05 |
| sjmc7 | these are search indexes | 17:06 |
| sjmc7 | not listener | 17:06 |
| sjmc7 | sorry - search aliases | 17:07 |
| RickA-HP | Ok. RT A listener has alias "aliasA1" pointing to index "index1". RT A search has alias "aliasA2" pointing to "index1". Same with RT B. Listener has alias "aliasB1" pointing to index "index1" and Search has alias "aliasB2" pointing to "index1". | 17:08 |
| RickA-HP | WHen re-syncing RT A, we update "aliasA1" (listener) to point to "index1" and the newly created index "index2". | 17:09 |
| RickA-HP | "aliasA2" still points to "index1". | 17:09 |
| sjmc7 | umm.. ok, although i don’t think it’s important to have separate aliases for listeners either. but for the sake of your example, yes | 17:09 |
| TravT | okay, for fun, i started drawing this | 17:10 |
| TravT | https://docs.google.com/presentation/d/1T35UXexvafxTaJ6Mw8JXrtSvmToZT5yqvufugsXxJ-4/edit?usp=sharing | 17:10 |
| RickA-HP | The updates from RT A go into both "index1" anmd "index2". The re-sync for RT A also goes into "index1" and "index2". | 17:10 |
| RickA-HP | RT A searches still go to "index1". | 17:11 |
| sjmc7 | ok | 17:11 |
| RickA-HP | Since RT B is associated with "index1", we update "aliasB1" (listener) to incldue "index1" and "index2" and start a resync. | 17:11 |
| sjmc7 | ok, so this is where my question 1 comes, but i’ll let you finish | 17:12 |
| RickA-HP | Now new updates from RT B go into "index1" and "index2" and the re-sync of RT B goes into "index1" and "index2". | 17:12 |
| RickA-HP | The RT A search is still going to "index1". | 17:12 |
| RickA-HP | When both RT A and RT B re-sync have compelted. We change "aliasA2" (Search) to "index2". | 17:12 |
| sjmc7 | ok | 17:12 |
| sjmc7 | this is where question 2 cdomes in | 17:13 |
| sjmc7 | btu carry on | 17:13 |
| RickA-HP | We finally switch "aliasA1" and "aliasB1" to "index2" only. | 17:13 |
| RickA-HP | Whew! | 17:13 |
| RickA-HP | Travt: Good job keeping up with the Google Docs! | 17:13 |
| RickA-HP | The assumption also is that the duplicate data already in index1 from the resync gets tossed by ES due to versioning. | 17:14 |
| sjmc7 | ok, so here’s question 1 | 17:15 |
| sjmc7 | what beenfit is there having separate aliases for A’s listener and B’s listener? why couldn’t we create index2 and the mappings, add index2 to a single alias, then have A and B start pushing data into it? | 17:16 |
| sjmc7 | if it’s easier we can do voice. i can update the spec review afterwards | 17:17 |
| RickA-HP | Logically, there is no benefit and "aliasA1" == "aliasB1". In terms of the current implementation does each listener store it's own index? | 17:17 |
| sjmc7 | i am enjoying watching travis draw though. it’s like the Bob Ross of Visio | 17:17 |
| RickA-HP | Travis doesn't have the hair for being Bob Ross. No on the other hand myself... | 17:18 |
| sjmc7 | each plugin has a configuration option for the index it uses | 17:18 |
| RickA-HP | So in my example, "aliasA1" is the variable within the plugin that stores the alias value. My assumption that values of the variables "aliasA1" and "aliasB1" would be the same. | 17:19 |
| RickA-HP | I actually diagrammed most of this in Visio on Tuesday. I will add it to the blueprint. | 17:20 |
| sjmc7 | so there is NOT an alias per resource type | 17:20 |
| sjmc7 | there’s an alias per index, that multiple RTs will use | 17:20 |
| RickA-HP | Yes. | 17:20 |
| RickA-HP | Sounds like the blueprint needs some nomenclature! | 17:21 |
| sjmc7 | well, you say that we have aliasA1 include index2, but that’s not correct | 17:21 |
| sjmc7 | we have the alias to which aliasA1 refers to include index2 | 17:21 |
| sjmc7 | the reason i’m asking is that you and travis both disagreed wiht me on the review, and i think it’s an important distinction | 17:22 |
| sjmc7 | and we need to get him one of those drawing tablets | 17:22 |
| RickA-HP | Yes, "aliasA1" the variable inside of the plug-in for RT A points to the physcial "aliasXX" in ES. The physical alias "aliasXX" contains a reference to "index1" | 17:23 |
| RickA-HP | The value of "aliasA1" should almost never change. | 17:24 |
| RickA-HP | Instead, we will change the physical ES alias "aliasXX". | 17:24 |
| RickA-HP | sjmc7: Do you still prefer a phone conferece or continue? | 17:25 |
| sjmc7 | no, i think we’re agreeing | 17:25 |
| TravT | hmm... so, rick, you are saying the alias in the config file is not an ES alias? | 17:25 |
| sjmc7 | it’ll be like now - the nova plugin and glance plugin both default to index_name=searchlight | 17:26 |
| TravT | let's do a hangout | 17:26 |
| sjmc7 | except it’ll be an alias, not a logical index | 17:26 |
| sjmc7 | ok | 17:26 |
| TravT | https://hangouts.google.com/call/gpavaaggm44jlsra3j6rhk2yima | 17:27 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 17:27 | |
| *** lakshmiS has joined #openstack-searchlight | 18:11 | |
| *** lakshmiS has quit IRC | 18:29 | |
| *** bpokorny has joined #openstack-searchlight | 18:44 | |
| *** lakshmiS has joined #openstack-searchlight | 18:51 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 19:02 | |
| *** TravT has quit IRC | 19:46 | |
| *** TravT has joined #openstack-searchlight | 19:56 | |
| *** TravT has quit IRC | 20:00 | |
| *** TravT has joined #openstack-searchlight | 20:14 | |
| TravT | sjmc7 lakshmiS RickA-HP (if you aren't too busy)... just looking at the final search query CLI patch | 21:04 |
| TravT | https://review.openstack.org/#/c/260899/ | 21:04 |
| TravT | i'm a little torn on whether this is an example like we spoke this morning where we could push it in and then iterate on formatting later | 21:05 |
| sjmc7 | with this one, to me it looks kind of broken | 21:05 |
| TravT | i don't want to lose the --details part | 21:05 |
| sjmc7 | the —detail output is terribly confusing | 21:05 |
| TravT | but the formatting is quite bad | 21:06 |
| sjmc7 | yeah, me neither | 21:06 |
| sjmc7 | maybe just a json blob? | 21:06 |
| sjmc7 | a massive json blob | 21:06 |
| sjmc7 | i don’t have a good answer, unfortunately. it’s sort of a problem inherent to unstructured search | 21:07 |
| sjmc7 | perhaps the plugin could define a list of fields? | 21:07 |
| sjmc7 | rather than the client? | 21:07 |
| TravT | still a lot of problematic maintenance | 21:07 |
| TravT | i don't have this solved on horizon side either | 21:07 |
| sjmc7 | yeah.. but at least it’s on the server side where the plugins live | 21:07 |
| sjmc7 | i agree it’s not ideal. i think having each client define it is a pain | 21:08 |
| TravT | but what does it define? | 21:09 |
| TravT | presentation is typically left to the presenter | 21:09 |
| sjmc7 | dunno. a list of ‘useful’ default fields | 21:09 |
| sjmc7 | yeah, i know :( | 21:09 |
| sjmc7 | maybe we do that then. have a list of mappings of type -> fields displayed | 21:09 |
| sjmc7 | back in a sec. this has been a problem since the early prototypes, how you use what is, to the client, unstructured data | 21:10 |
| lakshmiS | almost feel like having a cli interaction tool that can generate this command line for easy users | 21:10 |
| TravT | well and easier input is definitely a goal, but a a problem we have is the output right now. http://pasteboard.co/WM13caw.png | 21:12 |
| lakshmiS | how about a summary output and dig deeper with another query to get the details | 21:14 |
| RickA-HP | The output is not good. Cassandra has the same type of issue. At least the "pretty" option on ES helps. | 21:14 |
| TravT | let me grab a couple more example outputs to show you | 21:15 |
| RickA-HP | Could we have a "format" option to display the output in a clean way (like pretty=true in ES). But still retain the raw output. The raw output is good if trying to parse it with tools or in a pipe. | 21:16 |
| TravT | http://paste.openstack.org/show/484614/ | 21:17 |
| TravT | hmm, looks like other openstack client commands are pretty crappy too | 21:18 |
| TravT | openstack hypervisor list | 21:18 |
| TravT | then openstack hypervisor show | 21:18 |
| TravT | will provide copy paste | 21:19 |
| TravT | http://pasteboard.co/Y3tEI8r.png | 21:20 |
| lakshmiS | thats a comprise for showing the whole source output. alternatively we could provide a link which returns the source but that's one more hop to server | 21:20 |
| TravT | well see, we do provide simply output | 21:20 |
| TravT | openstack search query "" is simple and readable | 21:20 |
| TravT | adding --details makes it ugly | 21:20 |
| TravT | okay, this is already solved actually | 21:22 |
| sjmc7 | if you want the raw output, you’d probably use the python client itself | 21:22 |
| lakshmiS | well, --details is optional , so even though its ugly admin is asking for it | 21:22 |
| TravT | it is partially solved | 21:22 |
| TravT | osc supports max width | 21:23 |
| TravT | search query "" --details --max-width 50 | 21:23 |
| TravT | there is also a column and json format argument... | 21:23 |
| TravT | search query "" --details --max-width 50 -c Source -f json | 21:24 |
| TravT | kind og | 21:25 |
| TravT | kind of... | 21:25 |
| sjmc7 | —json kind of demonstrates the problem | 21:26 |
| sjmc7 | the json’s actually broken | 21:26 |
| sjmc7 | because of the comma->newline substitution | 21:26 |
| TravT | yeah... it isn't valid json | 21:28 |
| TravT | ttripp@ubuntu:~$ openstack search query "" --details -c Source -f value | python -m json.tool | 21:28 |
| TravT | Expecting property name: line 1 column 2 (char 1) | 21:28 |
| sjmc7 | something that’ll work, for now at least, is to have —details dump the entire _source out as unmodified json | 21:31 |
| TravT | yeah, so line 89 here | 21:32 |
| TravT | https://review.openstack.org/#/c/260899/7/searchlightclient/osc/v1/search.py | 21:32 |
| sjmc7 | yeah, and that’s the one i commented on i think, although i wasn’t forceful enough | 21:32 |
| TravT | let me see what that does locally | 21:32 |
| TravT | line 102 actually... | 21:35 |
| sjmc7 | it’s replacing all commas with newlines, and un-quoting (since json.dumps is putting quotes around everything) | 21:37 |
| TravT | well, that helped, but | 21:38 |
| TravT | this still doesn't work because the output doesn't provide a surrounding brace | 21:38 |
| sjmc7 | ? | 21:38 |
| TravT | ttripp@ubuntu:~$ openstack search query "name: cirros" --details -c Source -f value | python -m json.tool | 21:39 |
| TravT | Extra data: line 2 column 1 - line 5 column 1 (char 444 - 1876) | 21:39 |
| sjmc7 | the whole thing’s one giant string! | 21:40 |
| sjmc7 | take that function out entirely and it’ll do what you expect | 21:41 |
| TravT | yep | 21:43 |
| TravT | and i'd argue its fine | 21:43 |
| TravT | openstack search query "name: cirros" --details --max-width 50 | 21:43 |
| TravT | you can see it is json | 21:44 |
| TravT | so then you can do openstack search query "name: cirros" --details -c Source -f json | 21:44 |
| TravT | and it looks fine | 21:44 |
| sjmc7 | right, but the non-jsoned output looks a little weird | 21:44 |
| sjmc7 | having a look how this is done elsewhere | 21:45 |
| TravT | i think that is how it is done | 21:45 |
| sjmc7 | it’s sort of broken for glance image show, for instance; the properties list is output as a big string | 21:46 |
| sjmc7 | i.e. any formatting you do on column values applies to all formats | 21:47 |
| TravT | yeah same for hyervisor show | 21:48 |
| TravT | but that is a bit different | 21:48 |
| TravT | i mean this is a query | 21:48 |
| TravT | and -details gives you native result | 21:48 |
| TravT | the query is native | 21:48 |
| sjmc7 | right. what i am saying is, if you apply formatting to the _source column, it applies in both cases | 21:49 |
| sjmc7 | so if you turn it into a readable string in the table case, that’s how it’ll be presented in the json | 21:49 |
| sjmc7 | i’m not sure i’m making sense any more | 21:51 |
| * TravT trying to make brain work... just a sec | 21:52 | |
| TravT | got distracted | 21:52 |
| TravT | sjmc7: i know what you are saying | 21:53 |
| sjmc7 | that’s good. i’m not sure i do any more :) | 21:53 |
| TravT | i think it is better for it to output the json in table case | 21:53 |
| sjmc7 | what do we want to happen? | 21:53 |
| TravT | so that if you are using this in a pipeline of commands you can actually use the json in the -c (column only) case | 21:54 |
| sjmc7 | you mean in the —f json case? | 21:54 |
| TravT | yeah, what he said | 21:54 |
| sjmc7 | the -c just removes the id, score etc | 21:54 |
| TravT | well -c and -f case | 21:54 |
| sjmc7 | ok. in that case, we accept that in the tabular output, you’ll get a string representation of the _source field | 21:55 |
| sjmc7 | as in, str(_source) | 21:56 |
| sjmc7 | unicodes and all | 21:56 |
| TravT | better idea? | 21:57 |
| sjmc7 | the (only) alternative is to format _source in some way, and accept it’s not consumable with —json | 21:58 |
| TravT | maybe if the argument was --source (not --details), then it would make more sense | 21:59 |
| sjmc7 | yeah | 21:59 |
| TravT | you get a source column and it gives you source data that you can then operate on programatically | 21:59 |
| sjmc7 | ONLY in -f json mode | 21:59 |
| sjmc7 | not the tabular data | 21:59 |
| sjmc7 | i’m ok with that, i can’t think of a good way to get readable table and proper json format in the sonstraints of osc | 22:00 |
| sjmc7 | maybe we can get some input later | 22:00 |
| sjmc7 | i’ll put a review up | 22:00 |
| TravT | ok, i also have some comments with these commands in draft state | 22:01 |
| TravT | sjmc7: added some comment, please hop on with your own to +/- the suggestions. | 22:08 |
| sjmc7 | i did | 22:09 |
| sjmc7 | simultaneously | 22:09 |
| TravT | if --source optionally took arguments, they could be the fields to return... | 22:32 |
| TravT | e.g. --source returns all source | 22:32 |
| TravT | --source name,boo, it could return just those columns. | 22:33 |
| sjmc7 | yeah. maybe a future amendment? | 22:38 |
| TravT | yeah, not required for this patch | 22:55 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 23:34 | |
| *** bpokorny_ has joined #openstack-searchlight | 23:58 | |
| *** bpokorny_ has quit IRC | 23:58 | |
| *** bpokorny_ has joined #openstack-searchlight | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
