diff --git a/src/workbench/capabilities.rs b/src/workbench/capabilities.rs index caaca22..3f1ddce 100644 --- a/src/workbench/capabilities.rs +++ b/src/workbench/capabilities.rs @@ -48,8 +48,8 @@ pub fn probe_capabilities(params: &Value) -> Value { json!({"format": "docx", "available": false}), json!({"format": "ocr", "available": false}), // M-A0 freeze: format matrix rows remain unavailable until M-A3/M-A5. - json!({"format": "json", "available": false}), - json!({"format": "csv", "available": false}), + json!({"format": "json", "available": true}), + json!({"format": "csv", "available": true}), json!({"format": "image", "available": false}), json!({"format": "html", "available": false}), json!({"format": "svg", "available": false}), diff --git a/src/workbench/resource_view.rs b/src/workbench/resource_view.rs index 9081d28..aa42f6f 100644 --- a/src/workbench/resource_view.rs +++ b/src/workbench/resource_view.rs @@ -131,6 +131,19 @@ pub fn build_resource_view(req: &ResourceViewRequest) -> Result { &raw[..] }; let text = String::from_utf8_lossy(slice).into_owned(); + + // Structured JSON / CSV (M-A5 / Q-A5 subset): bounded parse, no exec. + if media == "application/json" { + return Ok(render_json_view( + &rel, &digest, &title, size, &text, truncated, req.max_bytes, kind, + )); + } + if media == "text/csv" { + return Ok(render_csv_view( + &rel, &digest, &title, size, &text, truncated, req.max_bytes, kind, + )); + } + let (renderer_id, format, fragment) = match media { "text/markdown" => ( "markdown", @@ -201,6 +214,10 @@ fn media_type_for(path: &str) -> &'static str { "application/x-marqdo" } else if lower.ends_with(".md") || lower.ends_with(".markdown") { "text/markdown" + } else if lower.ends_with(".json") { + "application/json" + } else if lower.ends_with(".csv") { + "text/csv" } else if lower.ends_with(".rs") || lower.ends_with(".ts") || lower.ends_with(".tsx") @@ -208,7 +225,6 @@ fn media_type_for(path: &str) -> &'static str { || lower.ends_with(".jsx") || lower.ends_with(".py") || lower.ends_with(".go") - || lower.ends_with(".json") || lower.ends_with(".toml") || lower.ends_with(".yaml") || lower.ends_with(".yml") @@ -229,6 +245,10 @@ fn kind_for(path: &str, media: &str) -> &'static str { "marqdo" } else if media == "text/markdown" { "markdown" + } else if media == "application/json" { + "json" + } else if media == "text/csv" { + "csv" } else if media == "text/x-code" { "code" } else if media.starts_with("text/") { @@ -238,6 +258,224 @@ fn kind_for(path: &str, media: &str) -> &'static str { } } +const JSON_OUTLINE_MAX: usize = 40; +const CSV_MAX_ROWS: usize = 200; +const CSV_MAX_COLS: usize = 32; + +fn render_json_view( + rel: &str, + digest: &str, + title: &str, + size: usize, + text: &str, + truncated: bool, + max_bytes: usize, + kind: &str, +) -> Value { + let mut diagnostics = Vec::new(); + if truncated { + diagnostics.push(json!({ + "code": "truncated", + "message": format!("preview truncated at {max_bytes} bytes") + })); + } + match serde_json::from_str::(text) { + Ok(val) => { + let outline = json_outline(&val, "", 0, &mut 0); + let pretty = serde_json::to_string_pretty(&val).unwrap_or_else(|_| text.to_string()); + let fragment = wrap_pre(&escape_html(&pretty)); + let completeness = if truncated { "partial" } else { "complete" }; + json!({ + "schema": "ResourceViewV1", + "schema_version": 1, + "resource": resource_ref(rel, "application/json", digest, title, size), + "renderer": {"id": "json", "version": "1", "trust": "sanitized"}, + "render": {"format": "sanitized_html", "fragment": fragment}, + "outline": outline, + "locators": [], + "completeness": completeness, + "no_exec": true, + "diagnostics": diagnostics, + "next_cursor": Value::Null, + "capabilities": {"read_only": true, "executed": false, "kind": kind} + }) + } + Err(e) => { + diagnostics.push(json!({ + "code": "json.parse_error", + "message": e.to_string() + })); + json!({ + "schema": "ResourceViewV1", + "schema_version": 1, + "resource": resource_ref(rel, "application/json", digest, title, size), + "renderer": {"id": "json", "version": "1", "trust": "sanitized"}, + "render": { + "format": "sanitized_html", + "fragment": wrap_pre(&escape_html(text)) + }, + "outline": [], + "locators": [], + "completeness": "partial", + "no_exec": true, + "diagnostics": diagnostics, + "next_cursor": Value::Null, + "capabilities": {"read_only": true, "executed": false, "kind": kind} + }) + } + } +} + +fn json_outline(val: &Value, path: &str, depth: usize, count: &mut usize) -> Vec { + if *count >= JSON_OUTLINE_MAX || depth > 3 { + return Vec::new(); + } + match val { + Value::Object(map) => { + let mut out = Vec::new(); + for (k, v) in map.iter().take(24) { + if *count >= JSON_OUTLINE_MAX { + break; + } + *count += 1; + let child = if path.is_empty() { + k.clone() + } else { + format!("{path}.{k}") + }; + let label = format!("{k}: {}", json_type_label(v)); + let mut node = json!({ + "id": child, + "label": label, + "kind": "json_key", + "locator": {"path": child} + }); + if depth < 2 { + let kids = json_outline(v, &child, depth + 1, count); + if !kids.is_empty() { + if let Some(obj) = node.as_object_mut() { + obj.insert("children".into(), json!(kids)); + } + } + } + out.push(node); + } + out + } + Value::Array(items) => { + *count += 1; + vec![json!({ + "id": path, + "label": format!("[{}] ({} items)", path, items.len()), + "kind": "json_array", + "locator": {"path": path} + })] + } + _ => Vec::new(), + } +} + +fn json_type_label(v: &Value) -> &'static str { + match v { + Value::Null => "null", + Value::Bool(_) => "bool", + Value::Number(_) => "number", + Value::String(_) => "string", + Value::Array(_) => "array", + Value::Object(_) => "object", + } +} + +fn render_csv_view( + rel: &str, + digest: &str, + title: &str, + size: usize, + text: &str, + truncated: bool, + max_bytes: usize, + kind: &str, +) -> Value { + let mut diagnostics = Vec::new(); + if truncated { + diagnostics.push(json!({ + "code": "truncated", + "message": format!("preview truncated at {max_bytes} bytes") + })); + } + let rows: Vec<&str> = text.lines().collect(); + let shown = rows.len().min(CSV_MAX_ROWS + 1); // header + data + if rows.len() > CSV_MAX_ROWS + 1 { + diagnostics.push(json!({ + "code": "csv.row_limit", + "message": format!("showing first {CSV_MAX_ROWS} data rows") + })); + } + let mut html = String::from("
"); + let mut locators = Vec::new(); + for (i, line) in rows.iter().take(shown).enumerate() { + let cells = split_csv_line(line); + let tag = if i == 0 { "th" } else { "td" }; + html.push_str(""); + for (c, cell) in cells.iter().take(CSV_MAX_COLS).enumerate() { + html.push('<'); + html.push_str(tag); + html.push('>'); + html.push_str(&escape_html(cell)); + html.push_str("'); + if i > 0 && c == 0 { + locators.push(json!({ + "id": format!("r{i}"), + "kind": "csv_row", + "locator": {"row": i, "col": c + 1} + })); + } + } + html.push_str(""); + } + html.push_str("
"); + let completeness = if truncated || rows.len() > CSV_MAX_ROWS + 1 { + "partial" + } else { + "complete" + }; + json!({ + "schema": "ResourceViewV1", + "schema_version": 1, + "resource": resource_ref(rel, "text/csv", digest, title, size), + "renderer": {"id": "csv", "version": "1", "trust": "sanitized"}, + "render": {"format": "sanitized_html", "fragment": html}, + "outline": [], + "locators": locators, + "completeness": completeness, + "no_exec": true, + "diagnostics": diagnostics, + "next_cursor": Value::Null, + "capabilities": {"read_only": true, "executed": false, "kind": kind} + }) +} + +/// Minimal CSV splitter (no quoted-newline support — preview only). +fn split_csv_line(line: &str) -> Vec { + let mut out = Vec::new(); + let mut cur = String::new(); + let mut in_q = false; + for ch in line.chars() { + match ch { + '"' => in_q = !in_q, + ',' if !in_q => { + out.push(cur.clone()); + cur.clear(); + } + c => cur.push(c), + } + } + out.push(cur); + out +} + fn wrap_pre(inner: &str) -> String { format!("
{inner}
") } @@ -394,4 +632,46 @@ mod tests { assert_eq!(errv["error"]["code"], "resource.path_escape"); let _ = fs::remove_dir_all(&dir); } + + #[test] + fn json_outline_and_no_exec() { + let dir = scratch_dir("json"); + let path = dir.join("cfg.json"); + fs::write(&path, r#"{"a":1,"b":{"c":true},"d":[1,2]}"#).unwrap(); + let req = ResourceViewRequest { + workspace_root: dir.clone(), + path: "cfg.json".into(), + revision: None, + max_bytes: 10_000, + }; + let view = build_resource_view(&req).unwrap(); + assert_eq!(view["no_exec"], true); + assert_eq!(view["renderer"]["id"], "json"); + assert_eq!(view["resource"]["media_type"], "application/json"); + assert!(view["outline"].as_array().unwrap().len() >= 2); + let frag = view["render"]["fragment"].as_str().unwrap(); + assert!(frag.contains(""a"") || frag.contains("\"a\"") || frag.contains("a")); + let _ = fs::remove_dir_all(&dir); + } + + #[test] + fn csv_table_bounded() { + let dir = scratch_dir("csv"); + let path = dir.join("t.csv"); + fs::write(&path, "name,age\nalice,3\nbob,4\n").unwrap(); + let req = ResourceViewRequest { + workspace_root: dir.clone(), + path: "t.csv".into(), + revision: None, + max_bytes: 10_000, + }; + let view = build_resource_view(&req).unwrap(); + assert_eq!(view["no_exec"], true); + assert_eq!(view["renderer"]["id"], "csv"); + let frag = view["render"]["fragment"].as_str().unwrap(); + assert!(frag.contains("