We will soon need `semantic_index` to be able to use
`CompletionProvider`. This is currently impossible due to a cyclic crate
dependency, because `CompletionProvider` lives in the `assistant` crate,
which depends on `semantic_index`.
This PR breaks the dependency cycle by extracting two crates out of
`assistant`: `language_model` and `completion`.
Only one piece of logic changed: [this
code](922fcaf5a6 (diff-3857b3707687a4d585f1200eec4c34a7a079eae8d303b4ce5b4fce46234ace9fR61-R69)).
* As of https://github.com/zed-industries/zed/pull/13276, whenever we
ask a given completion provider for its available models, OpenAI
providers would go and ask the global assistant settings whether the
user had configured an `available_models` setting, and if so, return
that.
* This PR changes it so that instead of eagerly asking the assistant
settings for this info (the new crate must not depend on `assistant`, or
else the dependency cycle would be back), OpenAI completion providers
now store the user-configured settings as part of their struct, and
whenever the settings change, we update the provider.
In theory, this change should not change user-visible behavior...but
since it's the only change in this large PR that's more than just moving
code around, I'm mentioning it here in case there's an unexpected
regression in practice! (cc @amtoaer in case you'd like to try out this
branch and verify that the feature is still working the way you expect.)
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>
Release Notes:
- Fixed a bug that caused focus to be lost when renames and inline
assists were scrolled offscreen.
---------
Co-authored-by: Nathan <nathan@zed.dev>
Note that this shouldn't have any visible user-facing behavior yet. The
feature is incomplete but we wanna merge early to avoid a long-running
branch.
Release Notes:
- N/A
---------
Co-authored-by: Nathan <nathan@zed.dev>
This PR updates the `/docs` slash command with the ability to
just-in-time index a package when there are not yet any results in the
index.
When running a `/docs` slash command, we fist check to see if there are
any results in the index that would match the search.
If there are, we go ahead and return them, as we do today.
However, if there are not yet any results we kick off an indexing task
as part of the command execution to fetch the results.
Release Notes:
- N/A
Previously, when using Vim mode and doing a line-wise visual selection,
the `assistant: quote selection` action would not work correctly, since
it would ignore that these were line-wise selections.
With this change, one can now select lines using visual line mode and
"quote selection works"
Release Notes:
- Fixes `assistant: quote selection` not working correctly when making
visual-line-mode selections in Vim mode.
This PR adds a streamlined way to consistently construct tab content for
items that only have textual content in the tabs.
The `Item` trait now has a new `tab_content_text` method that can be
used to return the textual content for the tab.
The `tab_content` method now has a default implementation that—unless
overridden—will construct a `Label` out of the text. This default
implementation also takes care of setting the label color based on the
active state of the tab, something that previously had to be repeated in
each `tab_content` implementation.
The majority of our tabs are now using `tab_content_text`.
Release Notes:
- N/A
Provide a current, broken state as an experimental way to browse
diagnostics.
The diagnostics are grouped by lines and reduced into a block that, in
case of multiple diagnostics per line, could be toggled back and forth
to show more diagnostics on the line.
Use `grouped_diagnostics::Deploy` to show the panel.
Issues remaining:
* panic on warnings toggle due to incorrect excerpt manipulation
* badly styled blocks
* no key bindings to navigate between blocks and toggle them
* overall odd usability gains for certain groups of people
Due to all above, the thing is feature-gated and not exposed to regular
people.
Release Notes:
- N/A
Release Notes:
- Added support for following into the assistant panel.
---------
Co-authored-by: Max <max@zed.dev>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Co-authored-by: Nathan <nathan@zed.dev>
This PR updates the `/docs` slash command with basic globbing support
for expanding docs.
A `*` can be added to the item path to signify the end of a prefix
match.
For example:
```
# This will match any documentation items starting with `auk::`.
# In this case, it will pull in the docs for each item in the crate.
/docs docs-rs auk::*
# This will match any documentation items starting with `auk::visitor::`,
# which will pull in docs for the `visitor` module.
/docs docs-rs auk::visitor::*
```
https://github.com/user-attachments/assets/5e1e21f1-241b-483f-9cd1-facc3aa76365
Release Notes:
- N/A
This PR updates the `/docs` slash command to show a warning to the user
if a crate's docs cannot be indexed due to the target directory not
containing docs:
<img width="782" alt="Screenshot 2024-07-11 at 5 11 46 PM"
src="https://github.com/user-attachments/assets/2f54f7a1-97f4-4d2d-b51f-57ba31e50a2f">
Release Notes:
- N/A
This PR updates the `/docs` slash command to use a more generic icon to
convey docs.
It was still using the Rust icon, a relic of when it was still
`/rustdoc`.
Release Notes:
- N/A
This pull request introduces collaboration for the assistant panel by
turning `Context` into a CRDT. `ContextStore` is responsible for sending
and applying operations, as well as synchronizing missed changes while
the connection was lost.
Contexts are shared on a per-project basis, and only the host can share
them for now. Shared contexts can be accessed via the `History` tab in
the assistant panel.
<img width="1819" alt="image"
src="https://github.com/zed-industries/zed/assets/482957/c7ae46d2-cde3-4b03-b74a-6e9b1555c154">
Please note that this doesn't implement following yet, which is
scheduled for a subsequent pull request.
Release Notes:
- N/A
This PR improves the completions for arguments in the `/docs` slash
command.
We achieved this by extending the `complete_argument` method on the
`SlashCommand` trait to return a `Vec<ArgumentCompletion>` instead of a
`Vec<String>`.
In addition to the completion `label`, `ArgumentCompletion` has two new
fields that are can be used to customize the completion behavior:
- `new_text`: The actual text that will be inserted when the completion
is accepted, which may be different from what is shown by the completion
label.
- `run_command`: Whether the command is run when the completion is
accepted. This can be set to `false` to allow accepting a completion
without running the command.
Release Notes:
- N/A
---------
Co-authored-by: Antonio <antonio@zed.dev>
This PR refactors the completion providers to only process a maximum
amount of completion requests at a time.
Also started refactoring language model providers to use traits, so it's
easier to allow specifying multiple providers in the future.
Release Notes:
- N/A
This PR adds a new `/docs` slash command to the Assistant. This slash
command replaces `/rustdoc`.
The `/docs` slash command works with different providers. There is
currently a built-in provider for rustdoc, but new providers can be
defined within extensions. The Gleam extension contains an example of
this.
When you first type `/docs` a completion menu will be shown with the
list of available providers:
https://github.com/zed-industries/zed/assets/1486634/32287000-5855-44d9-a2eb-569596f5abd9
After completing the provider you want to use then you can type the
package name and/or item path to search for the relevant docs:
https://github.com/zed-industries/zed/assets/1486634/6fc55a63-7fcd-42ea-80ce-08c670bf03fc
There are still some rough edges around completions that I would like to
get cleaned up in a future PR. Both of these seem to stem from the fact
that we're using an intermediate completion in the slash command:
1. Accepting a provider completion will show an error until you press
<kbd>Space</kbd> to continue typing.
- We need a way of not submitting a slash command when a completion is
accepted.
2. We currently need to show the provider name in the documentation item
completion list.
- Without it, the provider name gets wiped out when accepting a
completion, causing the slash command to become invalid.
Release Notes:
- N/A
This PR begins the process of making the backing infrastructure for the
`/rustdoc` command more generic such that it can be applied to
additional documentation providers.
In this PR we:
- Rename the `rustdoc` crate to `indexed_docs` as a more general-purpose
name
- Start moving rustdoc-specific functionality into
`indexed_docs::providers::rustdoc`
- Add an `IndexedDocsRegistry` to hold multiple `IndexedDocsStore`s (one
per provider)
We haven't yet removed the rustdoc-specific bits in the `DocsIndexer`.
That will follow soon.
Release Notes:
- N/A
This implements the functionality (paired with @as-cii), but we weren't
sure what the clearest name would be for the action. It's essentially
the inverse of "quote selection" - but what's the opposite of quoting
the selection?
One idea:
* Rename "quote selection" to "Insert **into** assistant"
* Name this "Insert **from** assistant"
Release Notes:
- Added action to insert from assistant into editor (default keybinding:
`cmd-<` on macOS, `ctrl-<` on Linux)
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Bennet <bennet@zed.dev>
Repeats project panel duplicated file name logic for prompts:
* add a ` copy` suffix
* if conflicts still, add a ` i` digit suffix, where `i` is the first
number available starting from 1
Release Notes:
- N/A
Release Notes:
- Fixed a bug that was causing system prompts to be dropped for
Anthropic models.
@JosephTLyons @notpeter We probably need to hot-fix this as I'm pretty
sure this affects the regular anthropic provider in addition to just the
feature-flagged cloud stuff. Wouldn't mind confirming that first so we
can communicate around it. 😬
This PR extracts a singular title bar (`title_bar::TitleBar`) from
`ui::TitleBar` and
`collab_ui::collab_titlebar_item::CollabTitlebarItem`.
This is a first step towards organizing title bar things into one place,
and standardizing platform titlebar/window control implementations.
Release Notes:
- N/A
This inline assistant prompt is one I designed that in my experience
works much better with Claude 3.5 Sonnet than the default prompt.
Mainly because it takes advantage of a weird property of our finetuning
which is that when you use XML tags it knows that it's doing a
machine-read tasks and stops trying to elide things for brevity. The
default prompt will often remove comments and otherwise add elisions for
brevity when doing large rewrites.
It also avoids giving the entire file content twice when the rewrite
region is large relative to the non-rewritten region.
Not necessarily meant to be merged as-is since it may mess up OAI
models. This is mainly meant for your reference. But everyone should be
using 3.5 Sonnet for coding use cases now anyhow 😛
Release Notes:
- N/A
TODO:
- [x] Moving the cursor out of the title editor should unselect any
selected text
Release Notes:
- N/A
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Richard <richard@zed.dev>
Fix a minor bug when the inline assistant model spits out an empty line
with leading indentation on it. This happens sometimes with Claude 3.5
Sonnet and currently it causes the following line to have the wrong
indentation.
Release Notes:
- N/A
