jupyter_protocol/media/mod.rs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646
//! Provides types and utilities for working with rich media content in Jupyter messages.
//!
//! This module defines the `Media` and `MediaType` structures, which represent
//! MIME bundles in Jupyter messages. These are used for rich content display
//! in notebooks and other Jupyter frontends.
//!
//! The main types in this module are:
//!
//! - [`Media`]: Represents a collection of media types.
//! - [`MediaType`]: An enum representing various MIME types.
//!
//! # Examples
//!
//! Creating a media bundle with multiple types:
//!
//! ```rust
//! use jupyter_protocol::media::{Media, MediaType};
//!
//! let media = Media::new(vec![
//! MediaType::Plain("Hello, world!".to_string()),
//! MediaType::Html("<h1>Hello, world!</h1>".to_string()),
//! ]);
//! ```
//!
//! Finding the richest media type:
//!
//! ```rust
//! use jupyter_protocol::media::{Media, MediaType};
//!
//! let media = Media::new(vec![
//! MediaType::Plain("Hello, world!".to_string()),
//! MediaType::Html("<h1>Hello, world!</h1>".to_string()),
//! MediaType::Markdown("**Hello, world!**".to_string()),
//! ]);
//!
//! let richest = media.richest(|media_type| match media_type {
//! MediaType::Html(_) => 3,
//! MediaType::Markdown(_) => 2,
//! MediaType::Plain(_) => 1,
//! _ => 0,
//! });
//!
//! assert!(matches!(richest, Some(MediaType::Html(_))));
//! ```
use serde::{de, Deserialize, Serialize};
use serde_json::Value;
use std::collections::HashMap;
pub mod datatable;
pub use datatable::TabularDataResource;
pub type JsonObject = serde_json::Map<String, serde_json::Value>;
/// An enumeration representing various Media types, otherwise known as MIME (Multipurpose Internet Mail Extensions) types.
/// These types are used to indicate the nature of the data in a rich content message such as `DisplayData`, `UpdateDisplayData`, and `ExecuteResult`.
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq)]
#[serde(rename_all = "snake_case")]
#[serde(tag = "type", content = "data")]
pub enum MediaType {
/// Plain text, typically representing unformatted text. (e.g. Python's `_repr_` or `_repr_pretty_` methods).
#[serde(rename = "text/plain")]
Plain(String),
/// HTML, (as displayed via Python's `_repr_html_` method).
#[serde(rename = "text/html")]
Html(String),
/// LaTeX, (as displayed using Python's `_repr_latex_` method).
#[serde(rename = "text/latex")]
Latex(String),
/// Raw JavaScript code.
#[serde(rename = "application/javascript")]
Javascript(String),
/// Markdown text, (as displayed using Python's `_repr_markdown_` method).
#[serde(rename = "text/markdown")]
Markdown(String),
/// SVG image text, (as displayed using Python's `_repr_svg_` method).
#[serde(rename = "image/svg+xml")]
Svg(String),
// Image data is all base64 encoded. These variants could all accept <Vec<u8>> as the
// data. However, not all users of this library will need immediate decoding of the data.
/// PNG image data.
#[serde(rename = "image/png")]
Png(String),
/// JPEG image data.
#[serde(rename = "image/jpeg")]
Jpeg(String),
/// GIF image data.
#[serde(rename = "image/gif")]
Gif(String),
/// Raw JSON Object
#[serde(rename = "application/json")]
Json(JsonObject),
/// GeoJSON data, a format for encoding a variety of geographic data structures.
#[serde(rename = "application/geo+json")]
GeoJson(JsonObject),
/// Data table in JSON format, requires both a `data` and `schema`.
/// Example: `{data: [{'ghost': true, 'says': "boo"}], schema: {fields: [{name: 'ghost', type: 'boolean'}, {name: 'says', type: 'string'}]}}`.
#[serde(rename = "application/vnd.dataresource+json")]
DataTable(Box<TabularDataResource>),
/// Plotly JSON Schema for for rendering graphs and charts.
#[serde(rename = "application/vnd.plotly.v1+json")]
Plotly(JsonObject),
/// Jupyter/IPython widget view in JSON format.
#[serde(rename = "application/vnd.jupyter.widget-view+json")]
WidgetView(JsonObject),
/// Jupyter/IPython widget state in JSON format.
#[serde(rename = "application/vnd.jupyter.widget-state+json")]
WidgetState(JsonObject),
/// VegaLite data in JSON format for version 2 visualizations.
#[serde(rename = "application/vnd.vegalite.v2+json")]
VegaLiteV2(JsonObject),
/// VegaLite data in JSON format for version 3 visualizations.
#[serde(rename = "application/vnd.vegalite.v3+json")]
VegaLiteV3(JsonObject),
/// VegaLite data in JSON format for version 4 visualizations.
#[serde(rename = "application/vnd.vegalite.v4+json")]
VegaLiteV4(JsonObject),
/// VegaLite data in JSON format for version 5 visualizations.
#[serde(rename = "application/vnd.vegalite.v5+json")]
VegaLiteV5(JsonObject),
/// VegaLite data in JSON format for version 6 visualizations.
#[serde(rename = "application/vnd.vegalite.v6+json")]
VegaLiteV6(JsonObject),
/// Vega data in JSON format for version 3 visualizations.
#[serde(rename = "application/vnd.vega.v3+json")]
VegaV3(JsonObject),
/// Vega data in JSON format for version 4 visualizations.
#[serde(rename = "application/vnd.vega.v4+json")]
VegaV4(JsonObject),
/// Vega data in JSON format for version 5 visualizations.
#[serde(rename = "application/vnd.vega.v5+json")]
VegaV5(JsonObject),
/// Represents Virtual DOM (nteract/vdom) data in JSON format.
#[serde(rename = "application/vdom.v1+json")]
Vdom(JsonObject),
// Catch all type for serde ease.
// TODO: Implement a custom deserializer so this extra type isn't in resulting serializations.
Other((String, Value)),
}
impl std::hash::Hash for MediaType {
fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
match &self {
MediaType::Plain(_) => "text/plain",
MediaType::Html(_) => "text/html",
MediaType::Latex(_) => "text/latex",
MediaType::Javascript(_) => "application/javascript",
MediaType::Markdown(_) => "text/markdown",
MediaType::Svg(_) => "image/svg+xml",
MediaType::Png(_) => "image/png",
MediaType::Jpeg(_) => "image/jpeg",
MediaType::Gif(_) => "image/gif",
MediaType::Json(_) => "application/json",
MediaType::GeoJson(_) => "application/geo+json",
MediaType::DataTable(_) => "application/vnd.dataresource+json",
MediaType::Plotly(_) => "application/vnd.plotly.v1+json",
MediaType::WidgetView(_) => "application/vnd.jupyter.widget-view+json",
MediaType::WidgetState(_) => "application/vnd.jupyter.widget-state+json",
MediaType::VegaLiteV2(_) => "application/vnd.vegalite.v2+json",
MediaType::VegaLiteV3(_) => "application/vnd.vegalite.v3+json",
MediaType::VegaLiteV4(_) => "application/vnd.vegalite.v4+json",
MediaType::VegaLiteV5(_) => "application/vnd.vegalite.v5+json",
MediaType::VegaLiteV6(_) => "application/vnd.vegalite.v6+json",
MediaType::VegaV3(_) => "application/vnd.vega.v3+json",
MediaType::VegaV4(_) => "application/vnd.vega.v4+json",
MediaType::VegaV5(_) => "application/vnd.vega.v5+json",
MediaType::Vdom(_) => "application/vdom.v1+json",
MediaType::Other((key, _)) => key.as_str(),
}
.hash(state)
}
}
/// A `Media` is a collection of data associated with different Media types.
/// It allows for the representation of rich content that can be displayed in multiple formats.
/// These are found in the `data` field of a `DisplayData` and `ExecuteResult` messages/output types.
///
#[derive(Default, Serialize, Deserialize, Debug, Clone)]
pub struct Media {
/// A map of Media types to their corresponding data, represented as JSON `Value`.
#[serde(
flatten,
deserialize_with = "deserialize_media",
serialize_with = "serialize_media_for_wire"
)]
pub content: Vec<MediaType>,
}
fn deserialize_media<'de, D>(deserializer: D) -> Result<Vec<MediaType>, D::Error>
where
D: serde::Deserializer<'de>,
{
// Jupyter protocol does pure Map<String, Value> for media types.
// Our deserializer goes a step further by having enums that have their data fully typed
let map: HashMap<String, Value> = HashMap::deserialize(deserializer)?;
let mut content = Vec::new();
for (key, value) in map {
// Check if the key matches ^application/(.*\\+)?json$ in order to skip the multiline string handling
if key.starts_with("application/") && key.ends_with("json") {
let media_type =
match serde_json::from_value(Value::Object(serde_json::Map::from_iter([
("type".to_string(), Value::String(key.clone())),
("data".to_string(), value.clone()),
]))) {
Ok(mediatype) => mediatype,
Err(_) => MediaType::Other((key, value)),
};
content.push(media_type);
continue;
}
// Now we know we're getting a plain string or an array of strings
let text: String = match value.clone() {
Value::String(s) => s,
Value::Array(arr) => arr
.into_iter()
.filter_map(|v| v.as_str().map(String::from))
.collect::<Vec<String>>()
.join(""),
_ => return Err(de::Error::custom("Invalid value for text-based media type")),
};
if key.starts_with("image/") {
// If we ever want to turn this into Vec<u8> we could do that here. We would need to strip all the whitespace from the base64
// encoded image too though. `let text = text.replace("\n", "").replace(" ", "");`
// For consistency with other notebook frontends though, we'll keep it the same
let mediatype: MediaType = match key.as_str() {
"image/png" => MediaType::Png(text),
"image/jpeg" => MediaType::Jpeg(text),
"image/gif" => MediaType::Gif(text),
_ => MediaType::Other((key.clone(), value)),
};
content.push(mediatype);
continue;
}
let mediatype: MediaType = match key.as_str() {
"text/plain" => MediaType::Plain(text),
"text/html" => MediaType::Html(text),
"text/latex" => MediaType::Latex(text),
"application/javascript" => MediaType::Javascript(text),
"text/markdown" => MediaType::Markdown(text),
"image/svg+xml" => MediaType::Svg(text),
// Keep unknown mediatypes exactly as they were
_ => MediaType::Other((key.clone(), value)),
};
content.push(mediatype);
}
Ok(content)
}
pub fn serialize_media_for_wire<S>(
content: &Vec<MediaType>,
serializer: S,
) -> Result<S::Ok, S::Error>
where
S: serde::Serializer,
{
serialize_media_with_options(content, serializer, false)
}
pub fn serialize_media_for_notebook<S>(media: &Media, serializer: S) -> Result<S::Ok, S::Error>
where
S: serde::Serializer,
{
serialize_media_with_options(&media.content, serializer, true)
}
pub fn serialize_media_with_options<S>(
content: &Vec<MediaType>,
serializer: S,
with_multiline: bool,
) -> Result<S::Ok, S::Error>
where
S: serde::Serializer,
{
let mut map = HashMap::new();
for media_type in content {
let (key, value) = match media_type {
MediaType::Plain(text)
| MediaType::Html(text)
| MediaType::Latex(text)
| MediaType::Javascript(text)
| MediaType::Markdown(text)
| MediaType::Svg(text) => {
let key = match media_type {
MediaType::Plain(_) => "text/plain",
MediaType::Html(_) => "text/html",
MediaType::Latex(_) => "text/latex",
MediaType::Javascript(_) => "application/javascript",
MediaType::Markdown(_) => "text/markdown",
MediaType::Svg(_) => "image/svg+xml",
_ => unreachable!(),
};
let value = if with_multiline {
let lines: Vec<&str> = text.lines().collect();
if lines.len() > 1 {
let entries = lines
.iter()
.map(|line| Value::String(format!("{}\n", line)));
Value::Array(entries.collect())
} else {
Value::Array(vec![Value::String(text.clone())])
}
} else {
Value::String(text.clone())
};
(key.to_string(), value)
}
// ** Treat images in a special way **
// Jupyter, in practice, will attempt to keep the multiline version of the image around if it was written in
// that way. We'd have to do extra tracking in order to keep this enum consistent, so this is an area
// where we may wish to diverge from practice (not protocol or schema, just practice).
//
// As an example, some frontends will convert images to base64 and then split them into 80 character chunks
// with newlines interspersed. We could perform the chunking but then in many cases we will no longer match.
MediaType::Jpeg(text) | MediaType::Png(text) | MediaType::Gif(text) => {
let key = match media_type {
MediaType::Jpeg(_) => "image/jpeg",
MediaType::Png(_) => "image/png",
MediaType::Gif(_) => "image/gif",
_ => unreachable!(),
};
let value = if with_multiline {
let lines: Vec<&str> = text.lines().collect();
if lines.len() > 1 {
let entries = lines
.iter()
.map(|line| Value::String(format!("{}\n", line)));
Value::Array(entries.collect())
} else {
Value::String(text.clone())
}
} else {
Value::String(text.clone())
};
(key.to_string(), value)
}
// Keep unknown media types as is
MediaType::Other((key, value)) => (key.clone(), value.clone()),
_ => {
let serialized =
serde_json::to_value(media_type).map_err(serde::ser::Error::custom)?;
if let Value::Object(obj) = serialized {
if let (Some(Value::String(key)), Some(data)) =
(obj.get("type"), obj.get("data"))
{
(key.clone(), data.clone())
} else {
continue;
}
} else {
continue;
}
}
};
map.insert(key, value);
}
map.serialize(serializer)
}
impl Media {
/// Find the richest media type in the bundle, based on the provided ranker function.
/// A rank of 0 indicates that the media type is not supported. Higher numbers indicate
/// that the media type is preferred over other media types.
///
/// # Examples
///
/// ```rust
/// use jupyter_protocol::media::{Media, MediaType};
///
/// let raw = r#"{
/// "text/plain": "FancyThing()",
/// "text/html": "<h1>Fancy!</h1>",
/// "application/json": {"fancy": true}
/// }"#;
///
/// let media: Media = serde_json::from_str(raw).unwrap();
///
/// let ranker = |media_type: &MediaType| match media_type {
/// MediaType::Html(_) => 3,
/// MediaType::Json(_) => 2,
/// MediaType::Plain(_) => 1,
/// _ => 0,
/// };
///
/// let richest = media.richest(ranker);
///
/// assert_eq!(
/// richest,
/// Some(MediaType::Html(String::from("<h1>Fancy!</h1>"))).as_ref()
/// );
///
/// ```
pub fn richest(&self, ranker: fn(&MediaType) -> usize) -> Option<&MediaType> {
self.content
.iter()
.filter_map(|mediatype| {
let rank = ranker(mediatype);
if rank > 0 {
Some((rank, mediatype))
} else {
None
}
})
.max_by_key(|(rank, _)| *rank)
.map(|(_, mediatype)| mediatype)
}
pub fn new(content: Vec<MediaType>) -> Self {
Self { content }
}
}
impl From<MediaType> for Media {
fn from(media_type: MediaType) -> Self {
Media {
content: vec![media_type],
}
}
}
impl From<Vec<MediaType>> for Media {
fn from(content: Vec<MediaType>) -> Self {
Media { content }
}
}
// Backwards compatibility with previous versions and Jupyter naming
pub type MimeBundle = Media;
pub type MimeType = MediaType;
#[cfg(test)]
mod test {
use datatable::TableSchemaField;
use serde_json::json;
use super::*;
#[test]
fn richest_middle() {
let raw = r#"{
"text/plain": "Hello, world!",
"text/html": "<h1>Hello, world!</h1>",
"application/json": {
"name": "John Doe",
"age": 30
},
"application/vnd.dataresource+json": {
"data": [
{"name": "Alice", "age": 25},
{"name": "Bob", "age": 35}
],
"schema": {
"fields": [
{"name": "name", "type": "string"},
{"name": "age", "type": "integer"}
]
}
},
"application/octet-stream": "Binary data"
}"#;
let bundle: Media = serde_json::from_str(raw).unwrap();
let ranker = |mediatype: &MediaType| match mediatype {
MediaType::Plain(_) => 1,
MediaType::Html(_) => 2,
_ => 0,
};
match bundle.richest(ranker) {
Some(MediaType::Html(data)) => assert_eq!(data, "<h1>Hello, world!</h1>"),
_ => panic!("Unexpected media type"),
}
}
#[test]
fn find_table() {
let raw = r#"{
"text/plain": "Hello, world!",
"text/html": "<h1>Hello, world!</h1>",
"application/json": {
"name": "John Doe",
"age": 30
},
"application/vnd.dataresource+json": {
"data": [
{"name": "Alice", "age": 25},
{"name": "Bob", "age": 35}
],
"schema": {
"fields": [
{"name": "name", "type": "string"},
{"name": "age", "type": "integer"}
]
}
},
"application/octet-stream": "Binary data"
}"#;
let bundle: Media = serde_json::from_str(raw).unwrap();
let ranker = |mediatype: &MediaType| match mediatype {
MediaType::Html(_) => 1,
MediaType::Json(_) => 2,
MediaType::DataTable(_) => 3,
_ => 0,
};
let richest = bundle.richest(ranker);
match richest {
Some(MediaType::DataTable(table)) => {
assert_eq!(
table.data,
Some(vec![
json!({"name": "Alice", "age": 25}),
json!({"name": "Bob", "age": 35})
])
);
assert_eq!(
table.schema.fields,
vec![
TableSchemaField {
name: "name".to_string(),
field_type: datatable::FieldType::String,
..Default::default()
},
TableSchemaField {
name: "age".to_string(),
field_type: datatable::FieldType::Integer,
..Default::default()
}
]
);
}
_ => panic!("Unexpected mime type"),
}
}
#[test]
fn find_nothing_and_be_happy() {
let raw = r#"{
"application/fancy": "Too ✨ Fancy ✨ for you!"
}"#;
let bundle: Media = serde_json::from_str(raw).unwrap();
let ranker = |mediatype: &MediaType| match mediatype {
MediaType::Html(_) => 1,
MediaType::Json(_) => 2,
MediaType::DataTable(_) => 3,
_ => 0,
};
let richest = bundle.richest(ranker);
assert_eq!(richest, None);
assert!(bundle.content.contains(&MediaType::Other((
"application/fancy".to_string(),
json!("Too ✨ Fancy ✨ for you!")
))));
}
#[test]
fn no_media_type_supported() {
let raw = r#"{
"text/plain": "Hello, world!",
"text/html": "<h1>Hello, world!</h1>",
"application/json": {
"name": "John Doe",
"age": 30
},
"application/vnd.dataresource+json": {
"data": [
{"name": "Alice", "age": 25},
{"name": "Bob", "age": 35}
],
"schema": {
"fields": [
{"name": "name", "type": "string"},
{"name": "age", "type": "integer"}
]
}
},
"application/octet-stream": "Binary data"
}"#;
let bundle: Media = serde_json::from_str(raw).unwrap();
let richest = bundle.richest(|_| 0);
assert_eq!(richest, None);
}
#[test]
fn ensure_array_of_text_processed() {
let raw = r#"{
"text/plain": ["Hello, world!"],
"text/html": "<h1>Hello, world!</h1>"
}"#;
let bundle: Media = serde_json::from_str(raw).unwrap();
assert_eq!(bundle.content.len(), 2);
assert!(bundle
.content
.contains(&MediaType::Plain("Hello, world!".to_string())));
assert!(bundle
.content
.contains(&MediaType::Html("<h1>Hello, world!</h1>".to_string())));
let raw = r#"{
"text/plain": ["Hello, world!\n", "Welcome to zombo.com"],
"text/html": ["<h1>\n", " Hello, world!\n", "</h1>"]
}"#;
let bundle: Media = serde_json::from_str(raw).unwrap();
assert_eq!(bundle.content.len(), 2);
assert!(bundle.content.contains(&MediaType::Plain(
"Hello, world!\nWelcome to zombo.com".to_string()
)));
assert!(bundle
.content
.contains(&MediaType::Html("<h1>\n Hello, world!\n</h1>".to_string())));
}
}