Partially migrate type answer to backend

Partially completes #1068, and will allow mobile clients to drop
their separate diff-match-patch imports. Does not yet try to handle
case folding or combining-char stripping, and leaves some of the outer
HTML wrapping up to the frontend for now.

The logic for rendering the provided string has changed: missing chars
are now only inserted if they follow a correct section, and the original
text is shown instead of hyphens. This is an experiment, and can be
changed if it's not well received.

Cargo.lock

@@ -55,6 +55,7 @@ dependencies = [
  "coarsetime",
  "criterion",
  "csv 1.1.6 (git+https://github.com/ankitects/rust-csv.git?rev=1c9d3aab6f79a7d815c69f925a46a4590c115f90)",
+ "dissimilar",
  "env_logger",
  "flate2",
  "fluent",
@@ -586,6 +587,12 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "dissimilar"
+version = "1.0.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8c97b9233581d84b8e1e689cdd3a47b6f69770084fc246e86a7f78b0d9c1d4a5"
+
 [[package]]
 name = "dtoa"
 version = "0.4.8"

cargo/BUILD.bazel

@@ -75,6 +75,15 @@ alias(
     ],
 )
 
+alias(
+    name = "dissimilar",
+    actual = "@raze__dissimilar__1_0_4//:dissimilar",
+    tags = [
+        "cargo-raze",
+        "manual",
+    ],
+)
+
 alias(
     name = "env_logger",
     actual = "@raze__env_logger__0_9_0//:env_logger",

cargo/crates.bzl

@@ -431,6 +431,16 @@ def raze_fetch_remote_crates():
         build_file = Label("//cargo/remote:BUILD.dirs-sys-next-0.1.2.bazel"),
     )
 
+    maybe(
+        http_archive,
+        name = "raze__dissimilar__1_0_4",
+        url = "https://crates.io/api/v1/crates/dissimilar/1.0.4/download",
+        type = "tar.gz",
+        sha256 = "8c97b9233581d84b8e1e689cdd3a47b6f69770084fc246e86a7f78b0d9c1d4a5",
+        strip_prefix = "dissimilar-1.0.4",
+        build_file = Label("//cargo/remote:BUILD.dissimilar-1.0.4.bazel"),
+    )
+
     maybe(
         http_archive,
         name = "raze__dtoa__0_4_8",

cargo/licenses.json

@@ -341,6 +341,15 @@
     "license_file": null,
     "description": "system-level helper functions for the dirs and directories crates"
   },
+  {
+    "name": "dissimilar",
+    "version": "1.0.4",
+    "authors": "David Tolnay <dtolnay@gmail.com>",
+    "repository": "https://github.com/dtolnay/dissimilar",
+    "license": "Apache-2.0",
+    "license_file": null,
+    "description": "Diff library with semantic cleanup, based on Google's diff-match-patch"
+  },
   {
     "name": "either",
     "version": "1.6.1",

cargo/remote/BUILD.dissimilar-1.0.4.bazel

@@ -0,0 +1,58 @@
+"""
+@generated
+cargo-raze crate build file.
+
+DO NOT EDIT! Replaced on runs of cargo-raze
+"""
+
+# buildifier: disable=load
+load("@bazel_skylib//lib:selects.bzl", "selects")
+
+# buildifier: disable=load
+load(
+    "@rules_rust//rust:defs.bzl",
+    "rust_binary",
+    "rust_library",
+    "rust_proc_macro",
+    "rust_test",
+)
+
+package(default_visibility = [
+    # Public for visibility by "@raze__crate__version//" targets.
+    #
+    # Prefer access through "//cargo", which limits external
+    # visibility to explicit Cargo.toml dependencies.
+    "//visibility:public",
+])
+
+licenses([
+    "notice",  # Apache-2.0 from expression "Apache-2.0"
+])
+
+# Generated Targets
+
+# Unsupported target "bench" with type "bench" omitted
+
+rust_library(
+    name = "dissimilar",
+    srcs = glob(["**/*.rs"]),
+    crate_features = [
+    ],
+    crate_root = "src/lib.rs",
+    data = [],
+    edition = "2018",
+    rustc_flags = [
+        "--cap-lints=allow",
+    ],
+    tags = [
+        "cargo-raze",
+        "crate-name=dissimilar",
+        "manual",
+    ],
+    version = "1.0.4",
+    # buildifier: leave-alone
+    deps = [
+    ],
+)
+
+# Unsupported target "test" with type "test" omitted

proto/anki/card_rendering.proto

@@ -26,6 +26,7 @@ service CardRenderingService {
   rpc EncodeIriPaths(generic.String) returns (generic.String);
   rpc DecodeIriPaths(generic.String) returns (generic.String);
   rpc StripHtml(StripHtmlRequest) returns (generic.String);
+  rpc CompareAnswer(CompareAnswerRequest) returns (generic.String);
 }
 
 message ExtractAVTagsRequest {
@@ -132,3 +133,8 @@ message StripHtmlRequest {
   string text = 1;
   Mode mode = 2;
 }
+
+message CompareAnswerRequest {
+  string expected = 1;
+  string provided = 2;
+}

pylib/anki/collection.py

@@ -1213,6 +1213,9 @@ class Collection(DeprecatedNamesMixin):
         "Not intended for public consumption at this time."
         return self._backend.render_markdown(markdown=text, sanitize=sanitize)
 
+    def compare_answer(self, expected: str, provided: str) -> str:
+        return self._backend.compare_answer(expected=expected, provided=provided)
+
     # Timeboxing
     ##########################################################################
     # fixme: there doesn't seem to be a good reason why this code is in main.py

pylib/rsbridge/cargo/BUILD.bazel

@@ -75,6 +75,15 @@ alias(
     ],
 )
 
+alias(
+    name = "dissimilar",
+    actual = "@raze__dissimilar__1_0_4//:dissimilar",
+    tags = [
+        "cargo-raze",
+        "manual",
+    ],
+)
+
 alias(
     name = "env_logger",
     actual = "@raze__env_logger__0_9_0//:env_logger",

qt/aqt/clayout.py

@@ -571,7 +571,7 @@ class CardLayout(QDialog):
         hadHR = origLen != len(txt)
 
         def answerRepl(match: Match) -> str:
-            res = self.mw.reviewer.correct("example", "sample")
+            res = self.mw.col.compare_answer("example", "sample")
             if hadHR:
                 res = f"<hr id=answer>{res}"
             return res

qt/aqt/reviewer.py

@@ -3,12 +3,9 @@
 
 from __future__ import annotations
 
-import difflib
-import html
 import json
 import random
 import re
-import unicodedata as ucd
 from dataclasses import dataclass
 from enum import Enum, auto
 from typing import Any, Callable, Literal, Match, Sequence, cast
@@ -24,7 +21,6 @@ from anki.scheduler.v3 import CardAnswer, NextStates, QueuedCards
 from anki.scheduler.v3 import Scheduler as V3Scheduler
 from anki.tags import MARKED_TAG
 from anki.types import assert_exhaustive
-from anki.utils import strip_html
 from aqt import AnkiQt, gui_hooks
 from aqt.browser.card_info import PreviousReviewerCardInfo, ReviewerCardInfo
 from aqt.deckoptions import confirm_deck_then_display_options
@@ -597,17 +593,10 @@ class Reviewer:
         buf = buf.replace("<hr id=answer>", "")
         hadHR = len(buf) != origSize
         # munge correct value
-        cor = self.mw.col.media.strip(self.typeCorrect)
+        expected = self.typeCorrect
-        cor = re.sub("(\n|<br ?/?>|</?div>)+", " ", cor)
+        provided = self.typedAnswer
-        cor = strip_html(cor)
-        # ensure we don't chomp multiple whitespace
-        cor = cor.replace(" ", "&nbsp;")
-        cor = html.unescape(cor)
-        cor = cor.replace("\xa0", " ")
-        cor = cor.strip()
-        given = self.typedAnswer
         # compare with typed answer
-        res = self.correct(given, cor, showBad=False)
+        output = self.mw.col.compare_answer(expected, provided)
         # and update the type answer area
         def repl(match: Match) -> str:
             # can't pass a string in directly, and can't use re.escape as it
@@ -616,7 +605,7 @@ class Reviewer:
 <span style="font-family: '{}'; font-size: {}px">{}</span>""".format(
                 self.typeFont,
                 self.typeSize,
-                res,
+                output,
             )
             if hadHR:
                 # a hack to ensure the q/a separator falls before the answer
@@ -644,84 +633,6 @@ class Reviewer:
             txt = ", ".join(matches)
         return txt
 
-    def tokenizeComparison(
-        self, given: str, correct: str
-    ) -> tuple[list[tuple[bool, str]], list[tuple[bool, str]]]:
-        # compare in NFC form so accents appear correct
-        given = ucd.normalize("NFC", given)
-        correct = ucd.normalize("NFC", correct)
-        s = difflib.SequenceMatcher(None, given, correct, autojunk=False)
-        givenElems: list[tuple[bool, str]] = []
-        correctElems: list[tuple[bool, str]] = []
-        givenPoint = 0
-        correctPoint = 0
-        offby = 0
-
-        def logBad(old: int, new: int, s: str, array: list[tuple[bool, str]]) -> None:
-            if old != new:
-                array.append((False, s[old:new]))
-
-        def logGood(
-            start: int, cnt: int, s: str, array: list[tuple[bool, str]]
-        ) -> None:
-            if cnt:
-                array.append((True, s[start : start + cnt]))
-
-        for x, y, cnt in s.get_matching_blocks():
-            # if anything was missed in correct, pad given
-            if cnt and y - offby > x:
-                givenElems.append((False, "-" * (y - x - offby)))
-                offby = y - x
-            # log any proceeding bad elems
-            logBad(givenPoint, x, given, givenElems)
-            logBad(correctPoint, y, correct, correctElems)
-            givenPoint = x + cnt
-            correctPoint = y + cnt
-            # log the match
-            logGood(x, cnt, given, givenElems)
-            logGood(y, cnt, correct, correctElems)
-        return givenElems, correctElems
-
-    def correct(self, given: str, correct: str, showBad: bool = True) -> str:
-        "Diff-corrects the typed-in answer."
-        givenElems, correctElems = self.tokenizeComparison(given, correct)
-
-        def good(s: str) -> str:
-            return f"<span class=typeGood>{html.escape(s)}</span>"
-
-        def bad(s: str) -> str:
-            return f"<span class=typeBad>{html.escape(s)}</span>"
-
-        def missed(s: str) -> str:
-            return f"<span class=typeMissed>{html.escape(s)}</span>"
-
-        if given == correct:
-            res = good(given)
-        else:
-            res = ""
-            for ok, txt in givenElems:
-                txt = self._noLoneMarks(txt)
-                if ok:
-                    res += good(txt)
-                else:
-                    res += bad(txt)
-            res += "<br><span id=typearrow>&darr;</span><br>"
-            for ok, txt in correctElems:
-                txt = self._noLoneMarks(txt)
-                if ok:
-                    res += good(txt)
-                else:
-                    res += missed(txt)
-        res = f"<div><code id=typeans>{res}</code></div>"
-        return res
-
-    def _noLoneMarks(self, s: str) -> str:
-        # ensure a combining character at the start does not join to
-        # previous text
-        if s and ucd.category(s[0]).startswith("M"):
-            return f"\xa0{s}"
-        return s
-
     def _getTypedAnswer(self) -> None:
         self.web.evalWithCallback("getTypedAnswer();", self._onTypedAnswer)
 

rslib/BUILD.bazel

@@ -77,6 +77,7 @@ rust_library(
         "//rslib/cargo:chrono",
         "//rslib/cargo:coarsetime",
         "//rslib/cargo:csv",
+        "//rslib/cargo:dissimilar",
         "//rslib/cargo:flate2",
         "//rslib/cargo:fluent",
         "//rslib/cargo:fnv",

rslib/Cargo.toml

@@ -101,3 +101,4 @@ id_tree = "1.8.0"
 zstd = { version="0.10.0", features=["zstdmt"] }
 num_cpus = "1.13.1"
 csv = { git="https://github.com/ankitects/rust-csv.git", rev="1c9d3aab6f79a7d815c69f925a46a4590c115f90" }
+dissimilar = "1.0.4"

rslib/cargo/BUILD.bazel

@@ -75,6 +75,15 @@ alias(
     ],
 )
 
+alias(
+    name = "dissimilar",
+    actual = "@raze__dissimilar__1_0_4//:dissimilar",
+    tags = [
+        "cargo-raze",
+        "manual",
+    ],
+)
+
 alias(
     name = "env_logger",
     actual = "@raze__env_logger__0_9_0//:env_logger",

rslib/i18n/cargo/BUILD.bazel

@@ -75,6 +75,15 @@ alias(
     ],
 )
 
+alias(
+    name = "dissimilar",
+    actual = "@raze__dissimilar__1_0_4//:dissimilar",
+    tags = [
+        "cargo-raze",
+        "manual",
+    ],
+)
+
 alias(
     name = "env_logger",
     actual = "@raze__env_logger__0_9_0//:env_logger",

rslib/i18n_helpers/cargo/BUILD.bazel

@@ -75,6 +75,15 @@ alias(
     ],
 )
 
+alias(
+    name = "dissimilar",
+    actual = "@raze__dissimilar__1_0_4//:dissimilar",
+    tags = [
+        "cargo-raze",
+        "manual",
+    ],
+)
+
 alias(
     name = "env_logger",
     actual = "@raze__env_logger__0_9_0//:env_logger",

rslib/linkchecker/cargo/BUILD.bazel

@@ -75,6 +75,15 @@ alias(
     ],
 )
 
+alias(
+    name = "dissimilar",
+    actual = "@raze__dissimilar__1_0_4//:dissimilar",
+    tags = [
+        "cargo-raze",
+        "manual",
+    ],
+)
+
 alias(
     name = "env_logger",
     actual = "@raze__env_logger__0_9_0//:env_logger",

rslib/src/backend/cardrendering.rs

@@ -15,6 +15,7 @@ use crate::{
         decode_iri_paths, encode_iri_paths, sanitize_html_no_images, strip_html,
         strip_html_preserving_media_filenames,
     },
+    typeanswer::compare_answer,
 };
 
 impl CardRenderingService for Backend {
@@ -147,6 +148,10 @@ impl CardRenderingService for Backend {
         .to_string()
         .into())
     }
+
+    fn compare_answer(&self, input: pb::CompareAnswerRequest) -> Result<pb::String> {
+        Ok(compare_answer(&input.expected, &input.provided).into())
+    }
 }
 
 fn rendered_nodes_to_proto(nodes: Vec<RenderedNode>) -> Vec<pb::RenderedTemplateNode> {

rslib/src/lib.rs

@@ -43,6 +43,7 @@ pub mod template_filters;
 pub(crate) mod tests;
 pub mod text;
 pub mod timestamp;
+mod typeanswer;
 pub mod types;
 pub mod undo;
 pub mod version;

rslib/src/typeanswer.rs

@@ -0,0 +1,233 @@
+// Copyright: Ankitects Pty Ltd and contributors
+// License: GNU AGPL, version 3 or later; http://www.gnu.org/licenses/agpl.html
+
+// FIXME: space to nbsp in output, or pre-wrap
+
+use std::borrow::Cow;
+
+use dissimilar::Chunk;
+use lazy_static::lazy_static;
+use regex::Regex;
+use unic_ucd_category::GeneralCategory;
+
+use crate::{
+    card_rendering::strip_av_tags,
+    text::{normalize_to_nfc, strip_html},
+};
+
+lazy_static! {
+    static ref LINEBREAKS: Regex = Regex::new(
+        r#"(?six)
+        (
+            \n
+            |
+            <br\s?/?>
+            |
+            </?div>
+        )+
+    "#
+    )
+    .unwrap();
+}
+
+struct DiffContext {
+    expected: String,
+    provided: String,
+}
+
+impl DiffContext {
+    fn new(expected: &str, provided: &str) -> Self {
+        DiffContext {
+            expected: prepare_expected(expected),
+            provided: prepare_provided(provided),
+        }
+    }
+
+    fn to_tokens(&self) -> DiffOutput<'_> {
+        let chunks = dissimilar::diff(&self.provided, &self.expected);
+        let mut provided = vec![];
+        let mut expected = vec![];
+        for chunk in chunks {
+            match chunk {
+                Chunk::Equal(text) => {
+                    provided.push(DiffToken {
+                        kind: DiffTokenKind::Good,
+                        text: text.into(),
+                    });
+                    expected.push(DiffToken {
+                        kind: DiffTokenKind::Good,
+                        text: text.into(),
+                    });
+                }
+                Chunk::Delete(text) => {
+                    provided.push(DiffToken {
+                        kind: DiffTokenKind::Bad,
+                        text: text.into(),
+                    });
+                }
+                Chunk::Insert(text) => {
+                    // If the proceeding text was correct, indicate text was missing
+                    if provided
+                        .last()
+                        .map(|v| v.kind == DiffTokenKind::Good)
+                        .unwrap_or_default()
+                    {
+                        provided.push(DiffToken {
+                            kind: DiffTokenKind::Missing,
+                            text: text.into(),
+                        });
+                    }
+                    expected.push(DiffToken {
+                        kind: DiffTokenKind::Missing,
+                        text: text.into(),
+                    });
+                }
+            }
+        }
+        DiffOutput { provided, expected }
+    }
+
+    fn to_html(&self) -> String {
+        let output = self.to_tokens();
+        let provided = render_tokens(&output.provided);
+        let expected = render_tokens(&output.expected);
+        format!(
+            "<div style='white-space: pre-wrap;'>{}</div>",
+            if no_mistakes(&output.expected) {
+                provided
+            } else {
+                format!("{provided}<br><span id=typearrow>&darr;</span><br>{expected}")
+            }
+        )
+    }
+}
+
+fn no_mistakes(tokens: &[DiffToken]) -> bool {
+    tokens.iter().all(|v| v.kind == DiffTokenKind::Good)
+}
+
+fn prepare_expected(expected: &str) -> String {
+    let without_av = strip_av_tags(expected);
+    let without_newlines = LINEBREAKS.replace_all(&without_av, " ");
+    let without_html = strip_html(&without_newlines);
+    normalize_to_nfc(&without_html).into()
+}
+
+fn prepare_provided(provided: &str) -> String {
+    normalize_to_nfc(provided).into()
+}
+
+#[derive(Debug, PartialEq)]
+enum DiffTokenKind {
+    Good,
+    Bad,
+    Missing,
+}
+
+#[derive(Debug, PartialEq)]
+struct DiffToken<'a> {
+    kind: DiffTokenKind,
+    text: Cow<'a, str>,
+}
+
+#[derive(Debug, PartialEq)]
+struct DiffOutput<'a> {
+    provided: Vec<DiffToken<'a>>,
+    expected: Vec<DiffToken<'a>>,
+}
+
+pub fn compare_answer(expected: &str, provided: &str) -> String {
+    DiffContext::new(expected, provided).to_html()
+}
+
+fn render_tokens(tokens: &[DiffToken]) -> String {
+    let text_tokens: Vec<_> = tokens
+        .iter()
+        .map(|token| {
+            let text = with_isolated_leading_mark(&token.text);
+            let encoded = htmlescape::encode_minimal(&text);
+            let class = match token.kind {
+                DiffTokenKind::Good => "typeGood",
+                DiffTokenKind::Bad => "typeBad",
+                DiffTokenKind::Missing => "typeMissed",
+            };
+            format!("<span class={class}>{encoded}</span>")
+        })
+        .collect();
+    text_tokens.join("")
+}
+
+/// If text begins with a mark character, prefix it with a non-breaking
+/// space to prevent the mark from joining to the previous token.
+fn with_isolated_leading_mark(text: &str) -> Cow<str> {
+    if let Some(ch) = text.chars().next() {
+        if GeneralCategory::of(ch).is_mark() {
+            return format!("\u{a0}{text}").into();
+        }
+    }
+    text.into()
+}
+
+#[cfg(test)]
+mod test {
+    use DiffTokenKind::*;
+
+    use super::*;
+
+    macro_rules! token {
+        ($kind:ident, $text:expr) => {
+            DiffToken {
+                kind: $kind,
+                text: $text.into(),
+            }
+        };
+    }
+
+    #[test]
+    fn tokens() {
+        let ctx = DiffContext::new("¿Y ahora qué vamos a hacer?", "y ahora qe vamosa hacer");
+        let output = ctx.to_tokens();
+        assert_eq!(
+            output.provided,
+            vec![
+                token!(Bad, "y"),
+                token!(Good, " ahora q"),
+                token!(Bad, "e"),
+                token!(Good, " vamos"),
+                token!(Missing, " "),
+                token!(Good, "a hacer"),
+                token!(Missing, "?"),
+            ]
+        );
+        assert_eq!(
+            output.expected,
+            vec![
+                token!(Missing, "¿Y"),
+                token!(Good, " ahora q"),
+                token!(Missing, "ué"),
+                token!(Good, " vamos"),
+                token!(Missing, " "),
+                token!(Good, "a hacer"),
+                token!(Missing, "?"),
+            ]
+        );
+    }
+
+    #[test]
+    fn html_and_media() {
+        let ctx = DiffContext::new("[sound:foo.mp3]<b>1</b> &nbsp;2", "1  2");
+        // the spacing is handled by wrapping html output in white-space: pre-wrap
+        assert_eq!(ctx.to_tokens().expected, &[token!(Good, "1  2")]);
+    }
+
+    #[test]
+    fn missed_chars_only_shown_in_provided_when_after_good() {
+        let ctx = DiffContext::new("1", "23");
+        assert_eq!(ctx.to_tokens().provided, &[token!(Bad, "23")]);
+        let ctx = DiffContext::new("12", "1");
+        assert_eq!(
+            ctx.to_tokens().provided,
+            &[token!(Good, "1"), token!(Missing, "2"),]
+        );
+    }
+}