Chinese Grammar Sentence Deck w/ Audio + Pinyin
A deck containing almost every example sentence from the AllSetLearning Chinese Grammar Wiki, all complete with pinyin, translation, usage guides and high-quality AzureTTS readings.
How to use this deck. This is primarily meant as a sentence deck, not a resource for grammar study, but it is viable to use it as the latter if you are diligent about reading the linked article for each card. The main purpose of this deck is to facilitate mostly-passive acquisition of vocabulary and grammatical structures.
Isn't a sentence-based grammar deck a strange concept? Sure, I guess. I mostly made this deck because I was curious about how well it would work in comparison to a standard grammar deck - for that I also made this other deck I made about japanese grammar. If you give them both a try, be sure to tell me which one feels more useful!
Licensing. All the content in the deck, in particular that taken from the wiki mentioned above, is CC BY-NC-SA 3.0.
Updates
1/4/25. Cleaned up card layout. Added formatting rules for night mode and mobile. Fixed some issues with weird sentence/pinyin spacing causing pauses in the tts. Modified card order, by hand-picking a few basic articles to put at the start of the deck (sentence order, shi, de, etc)
Notes
Order. All articles in the wiki are ordered by rank, and then alphabetically. All the example sentences from each article appear in the order that they appear in the article.
Sentence class. Sentences in the wiki are divided into those that have no class (most of them), those that have 'o' or 'x' class (right/wrong, for comparisons about common mistakes), and 'q', which is generally used for sentences that are not exactly wrong but have something off about them. 'q' and 'x' sentences are not included in this deck.
Dialogues. Some sentences are meant to be interpreted as part of a dialogue, s.t. reading a single sentence by itself leaves you without context. These are automatically detected and joined together into a single example sentence. All of them were checked by hand to avoid sentences being joined at the wrong place.
Usages. Most (~90%) of the sentences in the wiki are preceded by an 'usage' block that loosely describes how the grammatical structure being exemplified is meant to be used. Sometimes these usages are absent or in non-obvious order (e.g. usageA, usageB, exampleA, exampleB) - all of these patterns (examples without usages and usages without example) were detected by script and patched by hand.
Things done by hand. A few hours of work were invested in manually checking the data - some of this has been mentioned above, but also for miscellaneous checks, e.g. whenever latin characters show up in the chinese example sentences, or weird out-of-place html tags. In this sense the data should be relatively okay, but keep in mind that (a) I only checked the sentences flagged by the data extraction script (about 20% of them), and (b) I don't speak mandarin!!!
Tables. Some of the examples (or usages) on the wiki are not in example blocks but in tables, e.g. in the basic sentence order article. These were not collected.
Missing pinyin/trans. Some (about 5%) of the sentences in the wiki are missing pinyin or translation. For the sentences where this is missing, it is automatically generated, and a tag appears on their card indicating which parts of it were machine-generated.
HTML tags. All HTML tags are removed from sentences, most notably <em> and <strong> tags - so there is no emphasis on the sentences relating to the grammatical structure that they're supposed to exemplify.
Spacing. Sentences on the wiki have spaces between separate words, which is not how han characters are usually used. These spaces are removed on the front of the cards, but they are visible on the back of the cards.
Missing articles. All articles on the website were crawled. Only articles with the infobox describing their HSK level/keywords/etc (e.g. this one) had their example sentences extracted. A few of these articles (<5) were dropped in the manual checking phase because I saw no good way to make example sentences from them (e.g. the article about how numbers are structured, where the examples are a bunch of numbers).
Similar decks
There already exist on ankiweb a couple decks consisting primarily of content ripped from the AllSetLearning wiki. The three decks are generally similar, but with some important differences - explained below.
To Chinese Grammar (汉语 语法) HSK1 - HSK 6. This other deck expects more user input, in that the sentences have to be reconstructed from a bag of words, whereas this deck is meant as a pure recognition sort of thing. The TTS is very different in the two decks. Moreover there are some reports that some of the sentences on this deck are incorrect (this might have been fixed by now).
To Chinese Grammar Wiki Study Deck. This other deck is a lot closer to ours. Main differences are the inclusion of TTS and the fact that this deck also includes incorrect sentences (but expecting you to recognize them as such), whereas our deck is purely oriented to recognizing and understanding sentences, so we do not include the ones marked as incorrect.