This simplifies our glob filtering and makes sure Tables are always stored in a tree like structured apart from when returned from `FlattenTables`. This solves a lot of potential bugs and places where we need to do things like `if parent != nil`.

I think this also closes #475 as there is no need for another data structure.

The reasoning for this is that our tables are tree like structured and it's easier imo to always keep it like this everywhere across the code and not have multiple data structures with similar methods and so on. This required a small update to our filtering logic which just works as a DFS and in-place filtering (for that to work I had to add a copy method which I think can be  useful for other things in the future).

#### Summary

Fixes this panic:
```
12:22PM ERR column resolver finished with panic error={} client=pagerduty column=created_at duration=0.138344 module=pagerduty-src stack="failed to set column created_at: cannot parse  as Timestamptz\ngoroutine 402 [running]:\nruntime/debug.Stack()\n\t/usr/local/Cellar/go/1.19.1/libexec/src/runtime/debug/stack.go:24 +0x65\ngithub.com/cloudquery/plugin-sdk/plugins.(*SourcePlugin).resolveColumn.func1()\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/plugins/source_scheduler_dfs.go:226 +0x86\npanic({0x163c740, 0xc00049a860})\n\t/usr/local/Cellar/go/1.19.1/libexec/src/runtime/panic.go:884 +0x212\ngithub.com/cloudquery/plugin-sdk/schema.(*Resource).Set(0xc0005e8740?, {0x16f71e7, 0xa}, {0x15fb260?, 0xc0000e6158?})\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/schema/resource.go:74 +0x155\ngithub.com/cloudquery/plugin-sdk/schema.PathResolver.func1({0x18?, 0x1676020?}, {0xc0000be001?, 0xc0003a1a58?}, 0xc0005e8740, {{0x16f71e7, 0xa}, 0x9, {0x0, 0x0}, ...})\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/schema/resolvers.go:17 +0x73\ngithub.com/cloudquery/plugin-sdk/plugins.(*SourcePlugin).resolveColumn(0xc000294000, {0x17f5398, 0xc000516e40}, {{0x17f3878, 0xc00026a470}, 0x1, {0x0, 0x0}, {0xc000418a00, 0x5a, ...}, ...}, ...)\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/plugins/source_scheduler_dfs.go:233 +0x1cd\ngithub.com/cloudquery/plugin-sdk/plugins.(*SourcePlugin).resolveResource(0xc000294000, {0x17f53d0?, 0xc000319b60?}, 0xc0001f8f80, {0x17ef700, 0xc00034a7e0}, 0x14?, {0x16c49c0, 0xc0000e6100})\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/plugins/source_scheduler_dfs.go:209 +0x805\ngithub.com/cloudquery/plugin-sdk/plugins.(*SourcePlugin).resolveResourcesDfs.func1.1()\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/plugins/source_scheduler_dfs.go:148 +0x105\ncreated by github.com/cloudquery/plugin-sdk/plugins.(*SourcePlugin).resolveResourcesDfs.func1\n\t/Users/shimonp/projects/plugin-sdk-shimonp21/plugins/source_scheduler_dfs.go:144 +0xf6\n" table=pagerduty_user_notification_rules
```

When plugins have many tables, like the new Azure plugin we quickly run into memory issues with the current default.
This works on my machine ™️ with the new Azure plugin

---

Fixes a deadlock that occurs with very low concurrency settings. It was an off-by-one error 🤒

🤖 I have created a release *beep* *boop*
---


##
[1.12.1](v1.12.0...v1.12.1)
(2022-12-13)


### Bug Fixes

* Don't panic on empty-string for timestamp
([#489](#489))
([83813de](83813de))
* Fix deadlock off-by-one
([#493](#493))
([4ea9ed8](4ea9ed8))
* Reduce default concurrency
([#491](#491))
([f995da9](f995da9))
* Refactor glob filters
([#488](#488))
([cb5f6bb](cb5f6bb))

---
This PR was generated with [Release
Please](https://github.com/googleapis/release-please). See
[documentation](https://github.com/googleapis/release-please#release-please).