issue_comments

A start: ```diff diff --git a/datasette/utils/asgi.py b/datasette/utils/asgi.py index 8a2fa060..41ade961 100644 --- a/datasette/utils/asgi.py +++ b/datasette/utils/asgi.py @@ -118,7 +118,7 @@ class Request: return dict(parse_qsl(body.decode("utf-8"), keep_blank_values=True))

 @classmethod

• def fake(cls, path_with_query_string, method="GET", scheme="http", url_vars=None):
• def fake(cls, path_with_query_string, *, method="GET", scheme="http", url_vars=None): """Useful for constructing Request objects for tests""" path, _, query_string = path_with_query_string.partition("?") scope = { @@ -204,7 +204,7 @@ class AsgiWriter: )

-async def asgi_send_json(send, info, status=200, headers=None): +async def asgi_send_json(send, info, *, status=200, headers=None): headers = headers or {} await asgi_send( send, @@ -215,7 +215,7 @@ async def asgi_send_json(send, info, status=200, headers=None): )

-async def asgi_send_html(send, html, status=200, headers=None): +async def asgi_send_html(send, html, *, status=200, headers=None): headers = headers or {} await asgi_send( send, @@ -226,7 +226,7 @@ async def asgi_send_html(send, html, status=200, headers=None): )

-async def asgi_send_redirect(send, location, status=302): +async def asgi_send_redirect(send, location, *, status=302): await asgi_send( send, "", @@ -236,12 +236,12 @@ async def asgi_send_redirect(send, location, status=302): )

-async def asgi_send(send, content, status, headers=None, content_type="text/plain"): +async def asgi_send(send, content, status, *, headers=None, content_type="text/plain"): await asgi_start(send, status, headers, content_type) await send({"type": "http.response.body", "body": content.encode("utf-8")})

-async def asgi_start(send, status, headers=None, content_type="text/plain"): +async def asgi_start(send, status, *, headers=None, content_type="text/plain"): headers = headers or {} # Remove any existing content-type header headers = {k: v for k, v in headers.items() if k.lower() != "content-type"} @@ -259,7 +259,7 @@ async def asgi_start(send, status, headers=None, content_type="text/plain"):

async def asgi_send_file( - send, filepath, filename=None, content_type=None, chunk_size=4096, headers=None + send, filepath, filename=None, *, content_type=None, chunk_size=4096, headers=None ): headers = headers or {} if filename: @@ -284,7 +284,7 @@ async def asgi_send_file( )

-def asgi_static(root_path, chunk_size=4096, headers=None, content_type=None): +def asgi_static(root_path, *, chunk_size=4096, headers=None, content_type=None): root_path = Path(root_path)

 async def inner_static(request, send):

@@ -313,7 +313,7 @@ def asgi_static(root_path, chunk_size=4096, headers=None, content_type=None):

class Response: - def init(self, body=None, status=200, headers=None, content_type="text/plain"): + def init(self, body=None, , status=200, headers=None, content_type="text/plain"): self.body = body self.status = status self.headers = headers or {} @@ -346,6 +346,7 @@ class Response: self, key, value="", + , max_age=None, expires=None, path="/", @@ -374,7 +375,7 @@ class Response: self._set_cookie_headers.append(cookie.output(header="").strip())

 @classmethod

• def html(cls, body, status=200, headers=None):

• def html(cls, body, *, status=200, headers=None): return cls( body, status=status, @@ -383,7 +384,7 @@ class Response: )

  @classmethod - def text(cls, body, status=200, headers=None): + def text(cls, body, *, status=200, headers=None): return cls( str(body), status=status, @@ -392,7 +393,7 @@ class Response: )

  @classmethod - def json(cls, body, status=200, headers=None, default=None): + def json(cls, body, *, status=200, headers=None, default=None): return cls( json.dumps(body, default=default), status=status, @@ -401,7 +402,7 @@ class Response: )

  @classmethod - def redirect(cls, path, status=302, headers=None): + def redirect(cls, path, , status=302, headers=None): headers = headers or {} headers["Location"] = path return cls("", status=status, headers=headers) @@ -412,6 +413,7 @@ class AsgiFileDownload: self, filepath, filename=None, + , content_type="application/octet-stream", headers=None, ): diff diff --git a/datasette/app.py b/datasette/app.py index 03d1dacc..4d4e5584 100644 --- a/datasette/app.py +++ b/datasette/app.py @@ -190,6 +190,7 @@ class Datasette: def init( self, files=None, + *, immutables=None, cache_headers=True, cors=False, ```

Everything on https://docs.datasette.io/en/stable/internals.html that uses keyword arguments should do this I think.

This is a good idea - it's something I should do before Datasette 1.0.

I was a tiny bit worried about compatibility (Datasette is 3.7+) but it looks like they have been in Python since 3.0!

Demo: https://latest.datasette.io/_memory?sql=with+recursive+counter(x)+as+(%0D%0A++select+0%0D%0A++++union%0D%0A++select+x+%2B+1+from+counter%0D%0A)%2C%0D%0Ablah+as+(select++from+counter+limit+5000000)%0D%0Aselect+count()+from+blah

Implementing it like this, so at least you can copy and paste the SQL query back out again:

I'm not doing a full textarea because this error can be raised in multiple places, including on the table page itself. It's not just an error associated with the manual query page.

This is a good idea.

Another option here is to tie into a feature I built in sqlite-utils with this problem in mind but never introduced on the Datasette side of things: https://sqlite-utils.datasette.io/en/stable/python-api.html#cached-table-counts-using-triggers

That's a really interesting idea: for a lot of databases (those made out of straight imports from CSV) max(rowid) would indeed reflect the size of the table, but would be a MUCH faster operation than attempting a count(*).

That feels like a slightly different command to me - maybe sqlite-utils backup data.db data-backup.db? It doesn't have any of the mechanics for merging tables together. Could be a useful feature separately though.

• https://github.com/simonw/datasette/actions/runs/3131344150
• https://github.com/simonw/datasette/releases/tag/0.63a0
• https://pypi.org/project/datasette/0.63a0/

Tweeted about it too: https://twitter.com/simonw/status/1574481628507668480

New documentation: https://sqlite-utils.datasette.io/en/latest/contributing.html#using-just-and-pipenv

Documentation:

• https://sqlite-utils.datasette.io/en/latest/cli.html#cli-install
• https://sqlite-utils.datasette.io/en/latest/cli.html#cli-uninstall
• https://sqlite-utils.datasette.io/en/latest/cli-reference.html#install
• https://sqlite-utils.datasette.io/en/latest/cli-reference.html#uninstall

I tried the tips in https://stackoverflow.com/questions/15258831/how-to-handle-two-dashes-in-rest (not the settings change though, because I might want smart quotes elsewhere) and they didn't work.

Maybe I should disable smart quotes entirely?

I feel like there should be an escaping trick that works here though. I tried insert -\\-convert but it didn't help.

The implementation of this can be an almost exact copy of Datasette's, which was added in this commit: https://github.com/simonw/datasette/commit/01fe5b740171bfaea3752fc5754431dac53777e3

Current code for that is here: https://github.com/simonw/datasette/blob/0.62/datasette/cli.py#L319-L340 - which is improved to use the from runpy import run_module function.

This is also the kind of feature that would need to express itself in both the Python library and the CLI utility.

This is a really interesting idea.

I'm nervous about needing to set the rules for how duplicate tables should be merged though. This feels like a complex topic - one where there isn't necessarily an obviously "correct" way of doing it, but where different problems that people are solving might need different merging approaches.

Likewise, merging isn't just a database-to-database thing at that point - I could see a need for merging two tables using similar rules to those used for merging two databases.

So I think I'd want to have some good concrete use-cases in mind before trying to design how something like this should work. Will leave this thread open for people to drop those in!

This idea would make more sense if there was a good mechanism to say "run the conversion script held in this file" as opposed to passing it as an option. This would also make having to remember bash escaping rules (see tip) much easier!

shot-scraper has that for --javascript, using the --input option: https://shot-scraper.datasette.io/en/stable/javascript.html#shot-scraper-javascript-help

Maybe --convert-script would work here? Or --convert-file? It should accept - for stdin too.

Just saw your great write-up on this: https://jeqo.github.io/notes/2022-09-24-ingest-logs-sqlite/