* update api doc gen code

* update test

* update API docs format

* account for other terminal punctuation

* clarify api namespace target

* update schema test, add-period test

* update period test

* parens

* test endpoint-section

* cleanup

* Cam edits

Co-authored-by: Cam Saul <1455846+camsaul@users.noreply.github.com>

* trim whitespace

* file newlines

* Sort namespaces

Co-authored-by: Cam Saul <1455846+camsaul@users.noreply.github.com>
Co-authored-by: Cam Saul <github@camsaul.com>

* Add deny list of common passwords

* replace filtered pw list with full list

* switch to using a FileReader to check password line-by-line

* bypass deny list if complexity is set to easy

* fix tests

* more test fixes and remove unneeded ns

* fix typos

* fix reflection warning

* needed to import File

* fix final tests, and update docs to reflect change

* address initial review comments

* fix build after removing import statement

* address more review comments

* use InputStreamReader to read password file in JARs

* Update endpoint documentation

* Adding the API key documentation at the endpoint level

* Add all API changes

* Semantic types migration

* Fixup for basic querying

* Remove the relation type migration.

it makes the diff far too hard at the moment. need to be able to just
move everything from special_type -> semantic_type and then correct
the few that look at it to care about the effective type.

the PK/FK stuff gets really invasive and needs to be in its own much
smaller changeset

* Just rename special_type -> semantic_type for first change

* Special type -> semantic type

* special-type -> semantic-type

* SpecialType -> SemanticType

* special type -> semantic type

mostly in documentation, docstrings, etc

* Fix tests which relied on order of sets

database position was annotated by map-index'ing over the
set. changing `:special-type` to `:semantic-type` changed the order of
the seq produced from it.

* special -> semantic in schema_metadata

* Tim is awesome: Undo overeager special->semantic in docstrings

* Un-rename semantic_type in data_migrations

These migrations are run unless a migrations table marks them as
already having run. If they haven't run, then the db is older and the
column is special_type most likely. So we let them run as
`:special_type` and add some error handling to the migration runner
that is _opt in_.

```clojure
(defmigration ^{:author "camsaul", :added "0.20.0", :catch? true}
  migrate-field-types
  ...
  )

(try
  (@migration-var)
  (catch Exception e
    (if catch? ;; catch? from metadata
      (log/warn (format "Data migration %s failed: %s" migration-name (.getMessage e)))
      (throw e))))
```

* Fix merged master changes

just accepted their changes and fixed up rather than fix conflicts.

* Update API docs

* Fix spelling and rerun generator

* Fix linter

* Manually changing a bad generated section

This commit has the backend code needed to support alerts. From the
frontend perspective, Alerts is separate from pulses and has a
different endpoint, however much of the backend code and database
tables are shared. A few new columns were added to the pulse table to
support alerts, but most of the data is the same between pulses and
alerts.

This also adds several "meta-notifications" around alerts to let users
know when they've been added to an alert, removed from an alert or if
a question has changed such that the alert is no longer meaningful.

This fixes #1372, fixes #6261 and fixes #5768.