Fix CI test

Closes #6

See merge request !4

The test was written in a way that would work locally if you had
a specific S3 configuration, but failed in CI (or any other context with
a different set of values).

Add `POST /conversions` endpoint

Closes #4

See merge request !3

This endpoint allows a client to register a PDF that needs to be
converted to markdown.  It doesn't do the actual conversion, but instead
stores the request in a table which can then be used to drive batch
conversion actions.

Issue #4 Create endpoint to register a PDF / markdown conversion

We want our integration tests to run against a test database.  This
allows us to create more specific tests (e.g. predict ID values) in
addition to making sure we don't interact with real data.

I decided to use the PostgreSQL schema support instead of requiring
a separate `TEST_DATABASE_URL`.  This means there is less setup for
a dev environment but also makes it easier for us to drop / create the
schema as part of the tests.

There is a known deprecation issue and we don't need to be bothered by
it for now.  Eventually upstream will fix it in which case we should
remove the line that silences the warning.

This seems like standard boilerplate elsewhere, and was just omitted
initially.

I had thought this was enabled by default but it wasn't.  This just
means our imports will be sorted in a more standard / predictable way.

This also includes a new documented goal of staying on the most current
stable version of Python.

This seems like a good practice, but also we need some of the more
recent typing features.

This is basically the result of the initial alembic setup with the
addition of a single line to support custom schemas.  The reason for
schema support is that integration tests will be running against a test
schema.

We're going to need to use a database to store the state (and results)
of pdf conversions.

Add presigned post endpoint

Closes #2

See merge request !2

This endpoint returns an S3 presigned post, allowing a client to upload
a PDF to our S3 bucket.

Issue #2 Create endpoint to generate a presigned post

We'll be using `pytest` and `pytest-mock` for our unit and integration
tests.  This could have been set up earlier but I needed more runtime
context to know how the tools would fit together.

This is essentially a hello world application which just redirects users
to an empty docs page.  We'll start adding real endpoints over time,
don't worry.

Issue #2 Create endpoint to generate a presigned post

Set up base project

Closes #1

See merge request !1

This should run the linter as part of the CI process

Issue #1 Set up python project

Pyright is a type checking library for Python which should help catch
any type errors in the code base.

Issue #1 Setup python project

Ruff is a linter and formatter combo which can directly replace black,
flake8, pylint, and isort all in one swoop[1].

Issue #1 Setup python project
[1] https://docs.astral.sh/ruff/faq/

Issue #1 Set up python project

This is technically a file format used by pyenv, but even without pyenv
it's a handy way to signal what version of python to use.

Issue #1 Setup python project

Issue #1 Set up python project

Editorconfig is supported by a wide range of IDEs and helps editors use
proper whitespace rules before formatters are run.

Issue #1 Set up python project