Summary: TSIA. This snuck into D6741

Test Plan: NA

Reviewers: michelle, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6743

GitOrigin-RevId: 9e5041263f9814e0d429260cc54588beb47e990e

Summary: TSIA

Test Plan: `yarn dev` and storybook

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6741

GitOrigin-RevId: d41315c3dc384c1dd87aba23c791c3ce52572a64

Summary:
TSIA, this got a bit hairy because code-editor wanted access to
shortcuts to unregister monaco's bindings. I think I landed on a clean
solution but open to feedback.

Test Plan: `yarn dev` and storybook

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6737

GitOrigin-RevId: 4b2f3a9a2fa69a990aa55b51753c42d4b9474811

Summary: TSIA

Test Plan: `yarn dev` and storybook

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6735

GitOrigin-RevId: 30c3bb8c4ff19c2ba24002fa98a11dc0b612369d

Summary: TSIA

Test Plan: `yarn dev`

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6730

GitOrigin-RevId: 6b66dc48e603f1e571e8361f9b1ff5a566f5f264

Summary: TSIA

Test Plan: `yarn dev` and storybook

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6726

GitOrigin-RevId: 4076de63cb0cff74dcc6a67e102e39367c05535f

Summary:
TSIA.
`mui-theme` overrides the Palatte interface and a hence needed to
also be moved to ensure that things didn't complain about the Palette type and non
existent props. Sigh.

Test Plan: `yarn dev`

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6725

GitOrigin-RevId: 530872fdafc51b8959985c1f9279e3039321d223

Summary: TSIA

Test Plan: `yarn dev`

Reviewers: zasgar, michelle, nick, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6724

GitOrigin-RevId: 1b997c3f142e95ac438b21874c8aff6c5c17e834

Summary:
This sets up a pixie-components subpackage and moves the drawer compoenents
into it.

TODO: Fix styling. It looks like the styles for the drawer components gets overriden
when used by the pixie app. Should we be exporting unstyled components and let the pixie app
call the styling hooks?

TODO: Setup linting for components. The top level eslint + tsconfig doesn't play nicely
with a component with it's own eslint + tsconfig. This linter setup needs some help.

Stories are still in top level pixie to make it easier to debug whether the packaging
is working, to be moved into subcomponents eventually.

Test Plan: Check `yarn storybook` and `yarn dev`

Reviewers: zasgar, michelle, nick, philkuz, #engineering

Reviewed By: michelle, #engineering

JIRA Issues: PP-1680

Differential Revision: https://phab.corp.pixielabs.ai/D6710

GitOrigin-RevId: cc65338599b74809a5d23ccb1cb12145bea63418

Summary:
`go vet` assumes that colocated `.go` and `.cc` files must be built together
with `cgo`. In scenarios where we don't need/want to do so, let's move the `.go` source
into subfolders.

Note: This makes the build rule for the go libs looks a bit weird ... they end in `/go:go_default_library`
but IMO this is still cleaner than moving the `.cc` and `.h` files.

Test Plan: `bazel build //...` should succeed.

Reviewers: zasgar, michelle, philkuz, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6739

GitOrigin-RevId: c6c80acaa32d203aed6b59d88a20d2c8f50d33b6

Summary:
Also updated EqualsProto() to use DiffLines() to report the diffs.
A sample looks like this:

```
  deployment_spec {
    path: "src/stirling/obj_tools/testdata/dummy_go_binary_/dummy_go_binary"
  }
  structs {
l:  name: "out_table_value_t"
r:  name: "out_table_value2_t"
```

Test Plan: Jenkins

Reviewers: oazizi, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6711

GitOrigin-RevId: 0d6fb3b7faaa5e29e9d79cc79677de39df6855f5

Summary: they usually start in the non-hierarchical view by default. in some graphs, we always want to see them the hierarchical view first.

Test Plan: ran in webpack

Reviewers: zasgar, nserrino, nick, vihang, #engineering

Reviewed By: vihang, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6740

GitOrigin-RevId: 84fd29b087a9e164718188eb7f9cefdf04b99866

Summary:
This diff changes the behavior of px.sum on booleans to return a sum rather than a logical OR.

We need a consistent definition of what the sum of two booleans means. So I'm leaving this as a preview until we can make that call.

If we decide boolean + boolean = boolean, then this diff is not correct. Note that different languages (or even libraries within languages) have differing approaches.

The right long-term solution is to do this via data type casts. So this could be a temporary workaround until that is supported.

Test Plan: Modified the tests

Reviewers: philkuz, nserrino, #engineering, zasgar

Reviewed By: #engineering, zasgar

Differential Revision: https://phab.corp.pixielabs.ai/D6728

GitOrigin-RevId: 1b0c3a321276b1fb6871991d16d0421edbbec9f9

Summary: Was using this to debug the duplicate req/resp body bug.

Test Plan: Jenkins

Reviewers: oazizi, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6721

GitOrigin-RevId: 16d5261aed012605997b450cefbd218ee02643bd

Summary:
we should have a help icon in the sidebar which opens up the messenger.
when the user runs into an error, we should have a link that opens up the intercom messenger.

Test Plan:
gifs:
{F103061}

{F103062}

{F103063}

Reviewers: zasgar, nick, nserrino, #engineering, vihang

Reviewed By: #engineering, vihang

Subscribers: vihang

Differential Revision: https://phab.corp.pixielabs.ai/D6729

GitOrigin-RevId: b7d733961824d061e34720b7ae1c39072fc4ec12

Summary: we used to have this information in the signup event, but it has since disappeared because of some refactoring. it's helpful to have, for joining different tables in bigquery.

Test Plan: unit test

Reviewers: zasgar, nserrino, #engineering, vihang

Reviewed By: #engineering, vihang

Differential Revision: https://phab.corp.pixielabs.ai/D6731

GitOrigin-RevId: 7567c601fc4655362bcbf3f3f7f797766a482155

Summary: when adding users to the ui_prod table in segment, we also want to capture the emails so that we dont have to do extra joins to get that info in bigquery.

Test Plan: n/a

Reviewers: zasgar, nserrino, nick, vihang, #engineering

Reviewed By: vihang, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6732

GitOrigin-RevId: 31d52c0822b193928bef72e3eaa07355f154d98d

Summary: This reverts commit 01f74ebb2f608c65ddc0f6954d17c67c679a5190. Part of our decision to keep the time type around and instead make it more comp0atible with arrow.

Test Plan: teseted locally and things still work in normal mode.

Reviewers: nick, nserrino, vihang, zasgar, #engineering

Reviewed By: vihang, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6733

GitOrigin-RevId: 872e49dfc79c8c3d6775d3bdc83ef284fddf07b9

Summary: What I thought was a harmless Scout Rule change was, in fact, harmful.

Test Plan: Run `px/cluster` with a long start_time, like an hour. Cancel it before it returns. Then try to run `tcp_retransmits` with fresh tracepoints. The visual loading indicator should stick around for the correct duration. This time, also check that execution stats are retained.

Reviewers: michelle, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6734

GitOrigin-RevId: 0bb50d4e49cd8bee1b7dc5fb47a9c9edcc07c3d1

Summary: As every query run and retry cancels the previous one, the user cancelling one script then trying to run another that has mutations was triggering a second cancel on the first query. Due to awkward timing, that was setting loaded state to false a second time right after the second query started, so the tracepoints dialog thought it needed to hide. The fix is for the cancellation function to run once and replace itself with a noop.

Test Plan: Run `px/cluster` with a long start_time, like an hour. Cancel it before it returns. Then try to run `tcp_retransmits` with fresh tracepoints. The visual loading indicator should stick around for the correct duration.

Reviewers: nserrino, michelle, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6727

GitOrigin-RevId: e0e83aef3549c02676d6c462c5f5948c5ba23560

Summary: Before this change, a wrong type error would just yield "No arg value provided" instead should say that the type is wrong.

Test Plan: Tested on erroneous type and got the message out.

Reviewers: nick, vihang, michelle, #engineering

Reviewed By: michelle, #engineering

JIRA Issues: PP-2283

Differential Revision: https://phab.corp.pixielabs.ai/D6717

GitOrigin-RevId: 30016d26da8ba05ad90a4889224a6062fa41743e

Summary: Request building can create errors that were not properly handled. I moved the request builder in the observable function callback, but need some feedback on why we choose to send stuff around. A reviewer should help with that.

Test Plan: Tested with an erroneous vis spec where we had an integer instead of a string in the defaultValue and the error showed up, whereas the error would be emitted otherwise.

Reviewers: nick, vihang, michelle, #engineering

Reviewed By: nick, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6716

GitOrigin-RevId: bf79ab25eb0a164482f95f63b42a8af56fd544e4

Summary:
When new data enters a data table, if that table was scrolled to the top or bottom of its range, it sticks to wherever that edge moved. This is useful when a streaming query outputs an ever-flowing data set like HTTP requests and the user wants to watch that output live.

When scrolling to the top or bottom of a data table, stick to that edge if new data comes in.

This is especially useful for streaming queries, where the user may wish to monitor a live stream of recent events. Think `tail -f`. A future change will also put a finite scroll buffer on this for performance.
This is a first iteration. Depending on how this experience feels to users, we may try adding UI elements to enable/disable this mode directly or offer other methods.

Test Plan:
Try this streaming query. Sort it such that new data comes in either at the top or at the bottom, scroll to that edge, and release. The table should pin the scrollbar to that edge until you move it away.
```
import px
df = px.DataFrame(table='http_events', start_time='-2s')
df.pod = df.ctx['pod']
df = df.drop(['time_', 'upid', 'trace_role', 'http_minor_version', 'http_content_type', 'http_resp_message'])
px.display(df.stream())
```

Reviewers: nserrino, michelle, #engineering, vihang

Reviewed By: #engineering, vihang

Subscribers: philkuz, vihang

Differential Revision: https://phab.corp.pixielabs.ai/D6713

GitOrigin-RevId: 2cc67d26cf1df7aa467937a63d3a6a7aa0e010d1

Summary: Report whether a table is dynamic or static. Upstream this can be used in px/schemas one day to help filter out static vs dynamic tables.

Test Plan: Test case modified to add coverage

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

Subscribers: htroisi

Differential Revision: https://phab.corp.pixielabs.ai/D6718

GitOrigin-RevId: 25b4be0af6381dfb119518aa529bf21b98d7b983

Summary: Expand out the flags into its sub-fields, to make visual inspection and querying easier.

Test Plan: Tests modified.

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

Subscribers: htroisi

Differential Revision: https://phab.corp.pixielabs.ai/D6720

GitOrigin-RevId: 14deb5e63f81a0cec329d838d2bc83153c2c451e

Summary:
A version that has no external dependencies (no absl, no glogging). Also builds through makefile instead of Bazel. All to make it easy to run.

This is to be used for the blog.

Test Plan: Manually tested.

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6696

GitOrigin-RevId: 57189db1e11bdb210b2efa818bc16ccfdfab23e7

Summary: Depends on D6700, should land D6700 first.

Test Plan: Jenkins

Reviewers: oazizi, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6702

GitOrigin-RevId: e3a3a0d60fae25f133ccbca252e035bbb3ae51e9

Summary:
Add `react-is` to satisfy peer dependency for @storybook/addon-essentials
Upgrade `eslint-plugin-enzyme` to be compatible with our version of `eslint`

Test Plan: Jenkins + yarn build

Reviewers: zasgar, michelle, #engineering

Reviewed By: michelle, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6719

GitOrigin-RevId: ee50ed5da7353044064a7c2ed2ecee5a25143bc5

Summary:
This allows later refactoring of moving code generation for PerfBuffOutputAction
to code_gen.cc easier to review.

Test Plan: Jenkins

Reviewers: oazizi, #engineering

Reviewed By: oazizi, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6700

GitOrigin-RevId: b7af902695f4ace0431363c40e379546443a0e65

Summary:
we want to use react-cookie-banner to match what we will eventually use for our marketing site.
most of the code here is just styling, which we dont have a set mockup for, so feel free to nitpick the CSS

Test Plan:
screenshots:
{F101557}

{F101558}
updated zindex:
{F101792}

Reviewers: nick, zasgar, nserrino, #engineering

Reviewed By: nserrino, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6704

GitOrigin-RevId: f46804369136e3ae923a470aa56d700ea6bf81ac

Summary:
a huge portion of our bills are spent on stackdriver logging for our dev clusters, which most people don't actually look at.
updated the create_gke_cluster script to disable stackdriver logging by default.

Test Plan: created a new cluster, checked that the stackdriver logging was disabled

Reviewers: zasgar, nserrino, philkuz, #engineering, oazizi

Reviewed By: #engineering, oazizi

Differential Revision: https://phab.corp.pixielabs.ai/D6715

GitOrigin-RevId: 9ab941a928927b03382a9b32efa61c877346d4ce

Summary: To improve usability

Test Plan: Updated existing tests.

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

JIRA Issues: PP-2114

Differential Revision: https://phab.corp.pixielabs.ai/D6707

GitOrigin-RevId: 99b4ee4f7156e34626f06720b2db514b7155bd9c

Summary:
the Heap UDTFs raised the issue that we don't support ASID filters at the moment. This makes it so that we can filter if you use px.asid(). Not completely sure if I want to allow sweeping filter on asid. I'm worried about a situation like data w/ an asid column is stored on a Kelvin but was originally collected on a PEM. Filtering on an asid column in this case would mean we would prune hte Kelvin part which is wrong. I tried to make this as explicit as possible by only allowing
`px.asid() == 123` as the valid asid filter.

After a suggestion from Natalie, I've generalized the approach to create a Matcher factory that matches specific sub expressions. Users can add a matcher function to the metadata factory, the factory extracts the MatchExpr() static function. Whenever a match is made, factory creates the matcher operation and then Gets the Agents that should be pruned for the expression.

As before, we continue to support logical operators `and`/`or` on top of these fine-grained expressions.

Test Plan: Wrote a test for explicit case which this supports and against the implict case which this does not support.

Reviewers: nserrino, #engineering, vihang

Reviewed By: #engineering, vihang

Differential Revision: https://phab.corp.pixielabs.ai/D6647

GitOrigin-RevId: 8396edce0ff5f41c091503bb546ae1e0dc59bc99

Summary: Removing one piece of spaghetti reveals another...

Test Plan:
Try this streaming query. It should get results and still be cancellable.
```
import px
df = px.DataFrame(table='http_events', start_time='-2s')
df.pod = df.ctx['pod']
df = df.drop(['time_', 'upid', 'trace_role', 'http_minor_version', 'http_content_type', 'http_resp_message'])
px.display(df.stream())
```

Reviewers: nserrino, michelle, #engineering, philkuz, vihang

Reviewed By: #engineering, vihang

Differential Revision: https://phab.corp.pixielabs.ai/D6709

GitOrigin-RevId: c6df9dfa5289d1a5ef159f1150e95d06856ecf9b

Summary:
state setters allow you to use functions to set the state so you can use the current state to make the next state. When your state variable is a function, you must nest the function in another function

This is a missing fix I guess. Before this, couldnt' get any results on any scripts.

Test Plan: tested on prod and works.

Reviewers: nick, vihang, michelle, #engineering

Reviewed By: nick, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6708

GitOrigin-RevId: 8852fce748943e09b697003e8f724d80da11b4be

Summary: Part of a series of diffs that will be used to switch over to ST_TIME_NS isntead of using TIME64NS. The plan here was to support both the native data type and the semantic type simultaneously. In the future, we will deprecate the native data type.

Test Plan: Tested w/ a backend where time64ns is already switched over to INT64 + ST_TIME_NS and this wokred like a charm.

Reviewers: nick, vihang, michelle, nserrino, jamesbartlett, #engineering, zasgar

Reviewed By: #engineering, zasgar

Differential Revision: https://phab.corp.pixielabs.ai/D6706

GitOrigin-RevId: 01f74ebb2f608c65ddc0f6954d17c67c679a5190

Summary: UProbeTmpl is a concept only used by SocketTraceConnector, not by BCCWrapper at all, so it belongs with SocketTraceConnector.

Test Plan: Existing tests

Reviewers: yzhao, #engineering

Reviewed By: yzhao, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6698

GitOrigin-RevId: 1149e9bd872677f65009f4493b7b82ea98de8e3d

Summary:
1) Update MagicEnum and use its new built-in bazel rules.
2) Set defines to increase the max enum value that is supported.
3) Update the customization for SemanticType,

Test Plan: Jenkins

Reviewers: philkuz, jamesbartlett, #engineering, zasgar

Reviewed By: #engineering, zasgar

Differential Revision: https://phab.corp.pixielabs.ai/D6682

GitOrigin-RevId: f1c36b281de3237eac34b170f84dfb3a513f09e8

Summary:
Time fits semantic type dynamics better than primitive type. This change adds the semantic type to begin the switch.

The switch will work as follows:
1. Support ST_TIME_NS + INT64 in UI and CLI w/o deprecating TIME64NS
2. Replace all TIME64NS with ST_TIME_NS +INT64 in Vizier
3. Deprecate TIME64NS everywhere.

Test Plan: n/a doesn't change functionality

Reviewers: zasgar, oazizi, nserrino, michelle, #engineering, vihang

Reviewed By: #engineering, vihang

Subscribers: vihang

Differential Revision: https://phab.corp.pixielabs.ai/D6653

GitOrigin-RevId: df77f8403d3b8d88fdd24d86de7e6c54f96dc905

Summary: Doing this allows greater control over handling methods, reduced redundant re-renders, and an easier way to extend the possibilities going forward for things like window streaming. This comes at the cost of a more verbose listening model.

Test Plan: Actual behavior should be unaffected. Try a script that streams forever (so you can try canceling it), and try one that completes quickly. They should be identical to their production behavior.

Reviewers: nserrino, michelle, #engineering

Reviewed By: nserrino, #engineering

Differential Revision: https://phab.corp.pixielabs.ai/D6676

GitOrigin-RevId: e830cf87e5f3cdc9ff4b7aa6eea14abbc4e49ea0