This commit proposes the addition of "context serveres" and the
underlying protocol (model context protocol). Context servers allow
simple definition of slash commands in another language and running
local on the user machines. This aims to quickly prototype new commands,
and provide a way to add personal (or company wide) customizations to
the assistant panel, without having to maintain an extension. We can
use this to reuse our existing codebase, with authenticators, etc and
easily have it provide context into the assistant panel.
As such it occupies a different design space as extensions, which I
think are
more aimed towards long-term, well maintained pieces of code that can be
easily distributed.
It's implemented as a central crate for easy reusability across the
codebase
and to easily hook into the assistant panel at all points.
Design wise there are a few pieces:
1. client.rs: A simple JSON-RPC client talking over stdio to a spawned
server. This is
very close to how LSP work and likely there could be a combined client
down the line.
2. types.rs: Serialization and deserialization client for the underlying
model context protocol.
3. protocol.rs: Handling the session between client and server.
4. manager.rs: Manages settings and adding and deleting servers from a
central pool.
A server can be defined in the settings.json as:
```
"context_servers": [
{"id": "test", "executable": "python", "args": ["-m", "context_server"]
]
```
## Quick Example
A quick example of how a theoretical backend site can look like. With
roughly 100 lines
of code (nicely generated by Claude) and a bit of decorator magic (200
lines in total), one
can come up with a framework that makes it as easy as:
```python
@context_server.slash_command(name="rot13", description="Perform a rot13 transformation")
@context_server.argument(name="input", type=str, help="String to rot13")
async def rot13(input: str) -> str:
return ''.join(chr((ord(c) - 97 + 13) % 26 + 97) if c.isalpha() else c for c in echo.lower())
```
to define a new slash_command.
## Todo:
- Allow context servers to be defined in workspace settings.
- Allow passing env variables to context_servers
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>
This PR is just a refactor, to pave the way toward adding a view for
workflow step resolution. The entity carries the state of the tool
call's streaming output.
Release Notes:
- N/A
This PR is a refactor to pave the way for allowing the user to view and
edit workflow step resolutions. I've made tool calls work more like
normal streaming completions for all providers. The `use_any_tool`
method returns a stream of strings (which contain chunks of JSON). I've
also done some minor cleanup of language model providers in general,
removing the duplication around handling streaming responses.
Release Notes:
- N/A
Release Notes:
- Workflows can now create new files in folders that didn't exist
---------
Co-authored-by: jvmncs <7891333+jvmncs@users.noreply.github.com>
- Fix terminal slash command not working when terminal tab was placed in
center workspace
- Removed `--line-count` argument, you can now just pass a number to the
slash command
e.g. `/terminal 10` will show the last 10 lines of the active terminal
- Increase default context lines to 50
- We will revisit this once we add support for only including the last n
amount of commands that were run
Release Notes:
- N/A
This PR is a follow-up to
https://github.com/zed-industries/zed/pull/16216, as we want to run the
`/docs` command when completing a suggested package that has yet to be
indexed.
Release Notes:
- N/A
* renames `/tabs` to `/tab`
* allows to insert multiple tabs when fuzzy matching by the names
* improve slash command completion API, introduce a notion of multiple
arguments
* properly fire off commands on arguments' completions with
`run_command: true`
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <marshall@zed.dev>
This PR:
- Makes slash commands easier to compose by adding a concept,
`CompletionIntent`. When using `tab` on a completion in the assistant
panel, that completion item will be expanded but the associated command
will not be run. Using `enter` will still either run the completion item
or continue command composition as before.
- Fixes a bug where running `/diagnostics` on a project with no
diagnostics will delete the entire command, rather than rendering an
empty header.
- Improves the autocomplete rendering for files, showing when
directories are selected and re-arranging the results to have the file
name or trailing directory show first.
<img width="642" alt="Screenshot 2024-08-13 at 8 12 43 PM"
src="https://github.com/user-attachments/assets/97c96cd2-741f-4f15-ad03-7cf78129a71c">
Release Notes:
- N/A
This PR promotes package suggestions to a first-class concept on the
`IndexedDocsProvider` trait.
This will allow any implementer of `IndexedDocsProvider` to provide a
list of package names to suggest for use with `/docs`.
For the docs.rs provider we use the 250 most popular Rust crates (as
identified [here](https://lib.rs/std)), and for the rustdoc provider we
use the packages in the Cargo workspace.
Release Notes:
- N/A
/rant on
We have this issue where if a prompt starts with a slash command (e.g.
/workflow), the rendering is a bit messed up. The nested slash command
gets picked up as the parent of a command that includes it (/prompt).
This is due to how we parse slash commands; their output is obtained
asynchronously. When we run `/prompt "My prompt"` whose contents are
`/workflow`, we first include the prompt content verbatim and then
reparse the whole buffer, picking up /workflow as a new command (as if
it was typed by an user). The problem with that is that the range of
parent /prompt does not include the expanded range of a /workflow; in
fact, after doing full expansion of "My prompt", we lose track of the
parent-children relationship of these two slash commands and treat them
as if top-level user prompt was `/workflow/prompt "My prompt"` and not
`/prompt "My prompt"` (which, by the way, would not be parsable for us).
The "proper" fix would be to update the parent range whenever we parse a
new children within it. We could do that. But then, the question is;
what do we gain from it? Slash command output is put behind a crease,
which is fundamentally a fold. Given "My prompt", we'd have to put two
fold indicators on a single line even if the ranges were set up
correctly. So that merely moves the target elsewhere into yet another
issue. Even if we did solve two-fold problem somehow (by e.g. sorting
same-line folds by the end point), we would still be stuck with
suboptimal render. What do we gain from all that anyways? Proper
handling of a relatively obscure (although - at the same time - quite
common) edge case which may as well be handled by having /prompt insert
a new line if there's a slight chance that the edge case could occur.
And that hacky, "inproper" solution is what this PR does; in fact, it's
not the first time it was done, as /default also has the same issue
which it solves in precisely the same manner.
/rant off
Release Notes:
- N/A
For future reference: WIP branch of copy/pasting a mixture of images and
text: https://github.com/zed-industries/zed/tree/copy-paste-images -
we'll come back to that one after landing this one.
Release Notes:
- You can now paste images into the Assistant Panel to include them as
context. Currently works only on Mac, and with Anthropic models. Future
support is planned for more models, operating systems, and image
clipboard operations.
---------
Co-authored-by: Antonio <antonio@zed.dev>
Co-authored-by: Mikayla <mikayla@zed.dev>
Co-authored-by: Jason <jason@zed.dev>
Co-authored-by: Kyle <kylek@zed.dev>
This PR improves the initial experience of using `/docs docs-rs` with an
empty index.
We now show a brief explainer of what is expected (a crate name) as well
as list some popular Rust crates to try:
<img width="540" alt="Screenshot 2024-08-13 at 12 25 39 PM"
src="https://github.com/user-attachments/assets/942de250-2901-45df-9e3e-52ff3b3fc517">
Release Notes:
- N/A
Debug inspector broke immediately after merge as #16036 landed in
parallel; one of the changes of that PR is removing any steps whose
content was edited, which is what debug inspector happened to do. The
fix is to make the edit right past the step block.
Release Notes:
- N/A
Follow-up of https://github.com/zed-industries/zed/pull/16154
Reworks /tabs arguments to allow:
* current tab by default, if no arguments are present
* fuzzy-matching over paths of the related tabs
* `all` case to insert all tabs at once
Release Notes:
- N/A
Now, tabs have arguments, `active` (default, applied also for no
arguments case) and `all` to insert the active tab only or all tabs.
Release Notes:
- N/A
This adds a pulsating effect to the assistant header in case the message
is pending.
The pulsating effect is capped between 0.2 and 1.0 and I tried (with the
help of Claude) to give it a "breathing" effect, since I found the
normal bounce a bit too much.
Also opted for setting the `alpha` on the `LabelLike` things, vs.
overwriting the color, since I think that's cleaner instead of exposing
the color and mutating that.
https://github.com/user-attachments/assets/4a94a1c5-8dc7-4c40-b30f-d92d112db7b5
Release Notes:
- N/A
This PR simplifies our approach to indentation in the inline assistant
in hopes of improving our experience for Python. We tell the model to
generate the correct indentation in the prompt, and always start
generating at the start of the line. This may fall down for less capable
models, but I want to get a solid experience on the best models and then
figure the rest out later.
Also: We now prefer `./assets/prompts` as an overrides directory when
stdout is a PTY, so you can do `cargo run` and then iterate prompts for
the current run inside the current working copy.
cc @trishume @dsp-ant
Release Notes:
- Zed now allows the model to control indentation when performing inline
transformation. We're hoping this improves the indentation experience in
Python and other indentation-sensitive languages, but it does require
more from the model.
---------
Co-authored-by: Mikayla <mikayla@zed.dev>
This adds the requirement for users to accept the terms of service the
first time they send a message with the Cloud provider.
Once this is out and in a nightly, we need to add the check to the
server side too, to authenticate access to the models.
Demo:
https://github.com/user-attachments/assets/0edebf74-8120-4fa2-b801-bb76f04e8a17
Release Notes:
- N/A
Closes https://github.com/zed-industries/zed/issues/16042
This commit modifies the behavior of inline transformations to only
accept finished transformations when the user manually saves the file.
Previously, transformations were automatically accepted on any save
event, including autosaves.
This was achieved by updating the `Pane` and `Workspace` structs to emit
a new `UserSavedItem` event when a manual save occurs, and modifying the
`InlineAssistant` to register and handle this new event (instead of
`editor::Saved`).
Release Notes:
- N/A
This PR improves workflow step management and symbol matching. We've
optimized step pruning to remove any step that intersects an edit and
switched to normalized Levenshtein distance for more accurate symbol
matching.
Release Notes:
- N/A
This PR introduces a new script for iterative development of prompt
overrides in Zed.
Just `script/prompts link` and your running Zed should start using
prompts from `zed/assets/prompts`. Use `script/prompts unlink` to undo.
You can also link with `script/prompts link --worktree` to store the
prompts to a `../zed_prompts` worktree that's a sibling of your repo, in
case you don't want to mess with your working copy. Just don't forget
about it!
Key changes:
- Add new `script/prompts` for managing prompt overrides
- Rename `prompt_templates_dir` to `prompt_overrides_dir` for clarity
- Update paths to use `~/.config/zed/prompt_overrides` instead of
`~/.config/zed/prompts/templates`
- Adjust `PromptBuilder` to use the new `prompt_overrides_dir` function
These changes simplify the process of customizing prompts and provide a
more intuitive naming convention for override-related functionality.
Release Notes:
- N/A
We had to resort to this "hack" to get it to work, since nothing else we
tried (changing the `Pane`, changing the `ConfigurationView`, changing
the `AssistantPanel`, ...) worked.
Release Notes:
- N/A
Co-authored-by: Antonio <antonio@zed.dev>
Co-authored-by: Bennet <bennet@zed.dev>
Fixes#15923
Release Notes:
- Assistant workflow steps can now be applied and reverted directly from
within the assistant panel.
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Antonio <antonio@zed.dev>
