Make LanguageModel::use_any_tool return a stream of chunks (#16262)

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

crates/gpui/src/elements/img.rs

@@ -351,10 +351,13 @@ impl Asset for ImageAsset {
                     let mut body = Vec::new();
                     response.body_mut().read_to_end(&mut body).await?;
                     if !response.status().is_success() {
+                        let mut body = String::from_utf8_lossy(&body).into_owned();
+                        let first_line = body.lines().next().unwrap_or("").trim_end();
+                        body.truncate(first_line.len());
                         return Err(ImageCacheError::BadStatus {
                             uri,
                             status: response.status(),
-                            body: String::from_utf8_lossy(&body).into_owned(),
+                            body,
                         });
                     }
                     body