feat(hogql): sequential steps funnel

[thmsobrmlr]

Problem

Funnels need to be converted to HogQL (see the funnels section in PostHog/meta#130).

Changes

This PR:

• Implements all query steps necessary for a sequential (ordered) steps funnel without breakdowns and without converted/dropped people urls.
  • I tried to follow the existing implementation as closely as possible, since funnels span a large surface area.
  • Therefore I also left in a bunch of commented out lines to guide me when implementing the missing features.
  • These comments and the added TODOs grep TODO -rn posthog/hogql_queries/insights/funnels should be gone once the funnels rewrite is complete.
• Duplicates existing funnels tests for the HogQL variant:
  • posthog/queries/funnels/test/test_funnel.py
  • posthog/queries/funnels/test/conversion_time_cases.py
  • ee/clickhouse/queries/funnels/test/test_funnel.py (merged in with the regular tests)
  • I have removed 3 tests relating to single step funnels (those are not supported in the UI, so I haven't implemented them) and commentes out a test relating to validation errors (test_funnel_exclusions_invalid_params, to be implemented) and a failing cohort test (test_funnel_with_cohorts_step_filter, to be fixed).

Follow-ups

•  test_funnel_exclusions_invalid_params: implement validations (with Pydantic?)
• test_funnel_with_cohorts_step_filter: investigate this issue
• clean up all TODOs and commented out lines of code

How did you test this code?

I adapted the test script here https://gist.github.com/thmsobrmlr/ece8131eb7272893c274b3d9e77fc65c. It runs on most of my local insights, except some with validation errors and in one insight the conversion times don't match (to be investigated).

[github-actions]

Size Change: +85 B (0%)

Total Size: 2.22 MB

ℹ️ View Unchanged

Filename	Size	Change
frontend/dist/toolbar.js	2.22 MB	+85 B (0%)

_{compressed-size-action}

[mariusandra]

Wow, that's a lot of code. Reading through it all, it looks good, though obviously I didn't take the time to do a very deep review. Didn't spot anything odd either. Great work. 👍