-*- mode: org -*- #+TITLE: configuration git #+DESCRIPTION: documents - structuring, various output representations & search #+FILETAGS: :spine:hub: #+AUTHOR: Ralph Amissah #+EMAIL: [[mailto:ralph.amissah@gmail.com][ralph.amissah@gmail.com]] #+COPYRIGHT: Copyright (C) 2015 (continuously updated, current 2026) Ralph Amissah #+LANGUAGE: en #+STARTUP: content hideblocks hidestars noindent entitiespretty #+PROPERTY: header-args+ :eval never-export :exports code #+PROPERTY: header-args+ :noweb yes :padline no #+PROPERTY: header-args+ :results silent :cache no #+PROPERTY: header-args+ :mkdirp yes #+OPTIONS: H:3 num:nil toc:t \n:t ::t |:t ^:nil -:t f:t *:t * ./ ** README.md #+HEADER: :tangle ../README.md #+HEADER: :noweb yes #+BEGIN_SRC md # sisudoc-abstraction-format Specification of the **object-centric document abstraction** that [sisudoc-spine](https://git.sisudoc.org/sisudoc-spine) builds from SiSU markup, and of the two forms it is written in: **`.ssp`**, text and PEG-parsable, and **`.ocda.db`**, sqlite. The two carry one version number between them, because neither may express what the other cannot. ## What this is A specification, not a program. The repository ships: - [`ocda-model.md`](ocda-model.md) - **the model**: what an object is, the citation number, structure, sections, segments, digests. The part both serialisations share, and the place the three terms *ocda*, *abstraction* and `.ssp` are told apart. - [`ssp-format-specification.md`](ssp-format-specification.md) - **the `.ssp`**: the model written as text. - [`ocda-db-format-specification.md`](ocda-db-format-specification.md) - **the `.ocda.db`**: the model written as sqlite. - [`example-autonomous-contract.ssp`](example-autonomous-contract.ssp) - spine's output for that sample document, complete and unedited. - [`example-wealth-of-networks.ssp`](example-wealth-of-networks.ssp) - an excerpt of spine's output for that sample document: verbatim lines, marked elisions, keeping the preamble, the `@head` section and a run of body objects that includes an object embedding an image. The whole file runs to 27,769 lines. Both carry, in their opening comment lines, the command that regenerates them, so they can be diffed against a run rather than taken on trust. Both are output of **spine, format 2.1** (neither holds a poem, so from 2.0 to 2.1 each changed by its version line alone). ## Version The format is at **2.1**, and says so on the first line of every file: ``` % SiSU Document Abstraction v2.1 ``` The minor part is raised when a property is added, the major part when one changes meaning or goes away, and a reader checks the major part and refuses what it does not know. The history is in the specification. ## What is `.ssp`? Spine's processing pipeline goes ,**markup -> abstraction -> output**. The abstraction stage builds an in-memory object model: every paragraph, heading, table, footnote etc. is a numbered object with parent / sibling / type metadata ("OCN" - Object Citation Numbering). `.ssp` is a faithful text serialisation of that object model. It is designed for: - Inspecting and debugging spine's abstraction stage. - Diff-friendly version control of document structure (a `.ssp` diff between two markup edits shows exactly which objects changed). - Round-trip experiments: markup -> spine -> `.ssp` -> alternative renderers (HTML, LLM-ready chunks, ...) without rerunning spine. - A stable contract between spine and downstream consumers that do not want to link against spine's internal D types. Line types (full grammar in [`ssp-format-specification.md`](ssp-format-specification.md)): ``` @name { section / block open } section / block close [N] type ... object declaration (N = OCN, 0 if unnumbered) .name: value property of the preceding object | text content text line of the preceding object blank line (object separator) % comment comment (ignored by parser) ``` ## Audience - Developers writing readers / converters for spine's abstraction (readers written from this specification alone: sisudoc-ssp-reader, sisudoc-ocda-db-reader, now frozen with current readers offered and maintained within sisudoc-ocda-lib). - Anyone studying spine's internal document model without reading the D source. - LLM / search / indexing tooling that consumes structured document data downstream of spine. End-user document authors do not need to read this; they write `.sst` / `.ssm` markup, which spine then turns into `.ssp` along the way to the final output formats. ## Quickstart Read the spec: ```sh $EDITOR ssp-format-specification.md ``` Look at a real example: ```sh $EDITOR example-wealth-of-networks.ssp ``` Generate your own `.ssp` from a sample document via spine: ```sh # inside sisudoc-spine/ result/bin/spine --source --pod2 --output=/tmp/ssp path/to/pod ls /tmp/ssp/pod/*/media/abstraction/*.ssp ``` `--pod2` writes the abstraction as `.ssp` and bundles it into the source pod. (See sisudoc-spine's CLI help for the current flags.) ## License AGPL-3.0-or-later, matching the parent [sisudoc-spine](https://git.sisudoc.org/sisudoc-spine) project. ## Repository - **Canonical URL:** https://git.sisudoc.org/sisudoc-abstraction-format - **Clone:** `git clone --depth=1 git://git.sisudoc.org/specs/sisudoc-abstraction-format` ## Related projects - [sisudoc-spine](https://git.sisudoc.org/sisudoc-spine) - the document engine that writes `.ssp` and `.ocda.db` files. - [sisudoc-ocda-lib](https://git.sisudoc.org/sisudoc-ocda-lib) - the abstraction read as a library, from a .ssp, a .ocda.db or a .sisupod (D and Gleam). To be published. - [sisudoc-ocda-writer](https://git.sisudoc.org/sisudoc-ocda-writer) - sample writers over that library (text, markdown, html, epub, typst), each checked against spine. To be published. - [tree-sitter-sisu](https://git.sisudoc.org/tree-sitter-sisu) - grammar for the .sst / .ssm markup spine reads before writing a .ssp. #+END_SRC ** git .gitignore :gitignore: #+HEADER: :tangle "../.gitignore" #+BEGIN_SRC gitignore ,* !.gitignore !*.nix !flake.lock !*.md !*.ssp !org/ !**/*.org !**/COPYRIGHT #+END_SRC ** ocda-model.md #+HEADER: :tangle ../ocda-model.md #+BEGIN_SRC md # The object-centric document abstraction (OCDA) ,**The model.** What a document *is* once spine has parsed it, independent of any file it is written to. Two serialisations carry this model, and they carry **one version number between them**, because neither may express what the other cannot: | | | |----------------------------------------------------------------------|--------------------------------| | [`ssp-format-specification.md`](ssp-format-specification.md) | the OCDA as text, `.ssp` | | [`ocda-db-format-specification.md`](ocda-db-format-specification.md) | the OCDA as sqlite, `.ocda.db` | This document is what they share. It describes objects; they describe writing objects down. ## 1. Three words, and what each names The terms are used precisely in this repository and in spine, and the distinction is worth holding: | level | term | names | |--------------------|----------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------| | the model | **OCDA** | what a document is once parsed: numbered objects with structure, identity and digests. Independent of any file, any stage, any tool | | the stage | **abstraction** | the middle of spine's **markup -> abstraction -> output**: the act of building the model, and the point at which a document may now enter the pipeline | | the serialisations | **`.ssp`**, **`.ocda.db`** | the model written down, as text and as sqlite | A thing is named for the level it belongs to. Something true of the model whatever file it is in takes *ocda*; something about the pipeline stage takes *abstraction*; something that reads or writes one file form takes that form's name. ## 2. The object A document is a sequence of **objects**, grouped into sections. An object is a paragraph, a heading, a table, an endnote, a line of verse: the unit the document is made of and the unit it is cited by. Every object carries: - an **object citation number** (OCN), section 3; - an **identifier**, section 3.1; - a **kind** and a **type**, section 2.1; - its **place in the structure**: parent, ancestors, and for a heading its children and its last descendant, section 4; - the **segment** it is in, and for a heading an **anchor**, sections 7 and 7.1; - its **text**, in spine's internal representation; - a **digest** of that text. ### 2.1 Kind and type: two axes, not one An object is classified twice, and the two answer different questions. ,**Kind** (`is_a`) says *what the object is*, one of sixteen. Each kind belongs to one section, except `heading`, which appears in all nine: | kind | section | what it is | count | |----------------|--------------|--------------------------------------------------------------------------------|--------| | `para` | body | a paragraph, the ordinary unit of a document | 41,885 | | `heading` | *all nine* | a heading at one of the eight levels | 3,503 | | `toc` | toc | an entry in the generated table of contents | 3,215 | | `bookindex` | bookindex | an entry in the generated book index | 3,034 | | `endnote` | endnotes | a note, gathered from where it was called | 2,294 | | `code` | body | a code block | 1,623 | | `bibliography` | bibliography | a bibliography entry | 585 | | `comment` | body | markup comment, carried and not rendered | 450 | | `verse` | body | one verse of a poem | 151 | | `glossary` | glossary | a glossary entry | 56 | | `quote` | body | a block quotation | 53 | | `poem` | body | a poem: the container of its verses, OCN `0`, cited as their range (section 5) | 48 | | `block` | body | a block of grouped text | 45 | | `table` | body | a table, with its columns described | 32 | | `blurb` | blurb | a blurb entry | 28 | | `group` | body | a group block | 16 | ,**Type** (`is_of_type`) says *how the object's text behaves*, and takes only three values: | type | means | kinds | |-----------|---------------------------------------------------------|---------------------------------------------------------------------------------------| | `para` | flowing text, wrapped as the output sees fit | `para`, `heading`, `toc`, `bookindex`, `endnote`, `bibliography`, `glossary`, `blurb` | | `block` | text whose line breaks are significant and must be kept | `code`, `verse`, `quote`, `poem`, `block`, `table`, `group` | | `comment` | not rendered | `comment` | The mapping from kind to type is fixed: every `code` object is of type `block`, every `para` object of type `para`. **A renderer that has the type does not need to know all sixteen kinds** to decide whether it may reflow the text, which is what the second axis is for. ## 3. The object citation number The OCN is the heart of the model and the reason for it. ,**An object's OCN is the same in every format spine writes.** A paragraph that is object 42 in the HTML is object 42 in the EPUB, in the PDF, in the plain text, in the search database and in both serialisations of the abstraction. A citation to object 42 of a document is therefore a citation to a fixed piece of text, not to a page in one edition or a scroll position in one rendering. Three properties follow, and each is a trap for a reader that assumes otherwise: - **Not every object has one.** Generated furniture, and objects the markup has suppressed, carry `0`. There are many of them: 7,543 across the 36 document reference collection. - **`0` is therefore not an identity.** Objects sharing `0` are not the same object, and `0` must never be used as a key. - **A non-zero OCN is unique within a document**, from format 2.1: checked over the reference collection, no non-zero OCN occurs twice in any of the 36 document-languages. **Before 2.1 it was not**: a poem carried the OCN of its last verse, so the same number named two objects (48 such pairs in the collection, 21 in `through_the_looking_glass`). A reader of an older artefact keys on position within a section when one row is wanted. ### 3.1 The identifier Every object also carries an **identifier**: the name by which an output refers to it. An HTML anchor, a LaTeX `\hypertarget`, a link from a book index entry to the object it indexes, are all written from the identifier. ,**The identifier of an object with a non-zero OCN is its OCN**, written as text. That is the ordinary case and it is the reason a citation to object 42 can be turned into a link without consulting anything. An object with OCN `0` has no OCN to take an identifier from, so one is assigned: `a1`, `a2`, `a3`, ... in document order over the objects that need one. There are few of them, eleven in a book, and they are the generated furniture of section 3. ,**A reader must arrive at the identifier whether or not the serialisation wrote it down**; see section 9. ## 4. Levels A document has **eight heading levels**, in two groups of four: | level | index | what it is | |-------|-------|----------------------------------------------| | `A` | 0 | the document title. Exactly one per document | | `B` | 1 | | | `C` | 2 | the title matter below it | | `D` | 3 | | | `1` | 4 | the top level of the document body | | `2` | 5 | | | `3` | 6 | below that | | `4` | 7 | | `A` to `D` are the title levels, `1` to `4` the body levels, and the index is what the eight-slot arrays of section 5 are indexed by. A heading that claims no level has 9, which is the "not a heading level" value and is why `.heading_lev_collapsed` is emitted only when it is below 9. ,**Collapsed levels.** A document that uses `A` and then `1`, skipping `B`, `C` and `D`, has a gap in its levels that is an artefact of the markup rather than of the document. Every level-indexed array therefore comes in two forms: as marked up, and **collapsed**, with the unused levels removed so that the remaining ones are contiguous. Outputs that care about nesting depth use the collapsed form; anything tracing back to the markup uses the other. ## 5. Structure Structure is carried by number, not by nesting: an object names its parent rather than containing its children. This is what lets the serialisations be flat and the model be walked in either direction. - **`parent`** - the OCN of the object this one hangs from. **A parent of `0` is not "no parent"**: it is a parent heading the author has made non-substantive, its OCN suppressed or a dummy heading introduced to segment an output. The distinction is preserved deliberately, and a reader walking upward must stop at `0` rather than treat it as missing, or the walk will not terminate. - **`ancestors`** - **eight slots, indexed by level** (section 4): the OCN of the heading at each level above this object, `0` where there is none, so the path to the root is available without walking it. **A heading's own OCN sits in its own level's slot**: checked over the reference collection, 3,432 headings of 3,432. So for a heading the array is its path including itself, and for any other object it is the path to the heading above it. - **`ancestors_collapsed`** - the same, in collapsed levels. - **`children`** - a heading's child headings. - **`last_descendant`** - the last OCN under a heading, which makes a subtree a **range** rather than a traversal: everything from the heading's OCN to its last descendant is inside it. - **`ocn_range`** - for a **container**, the first and last OCN of what it holds (2.1). A poem is the one container: its verses are the citable units, one object each, and the poem is not; it carries OCN `0` and no identifier, **comes immediately before its first verse**, and is cited as the range of them (`173-180`). So a writer that meets a poem knows what follows and where it ends, and can wrap the verses in one pass. The poem has no text of its own. `0 0`, or absent, where the object is not a container. **Before 2.1** the poem came *after* its last verse, carried that verse's OCN and identifier, and stated no range. ## 6. Sections Objects are grouped into nine sections, in this order: ``` head toc body endnotes glossary bibliography bookindex blurb tail ``` `head` is the document's title matter, `body` the document proper, and the rest are generated or gathered. `tail` is the end of document some output writers use. A section with no objects is not carried. ## 7. Segments An output may be one file or many. A **segment** is one of the many: a chapter-sized piece of HTML or EPUB with its own name and its own navigation. Two different questions are asked about segments, and the model answers both separately because the answers differ: - **which segment does this heading open?** Set on headings only. - **which segment is this object in?** Set on nearly every object. Without the second, a reader cannot say where a paragraph belongs. The search database reads the second. ### 7.1 The anchor A heading carries an **anchor**: the name an output links to in order to arrive at that heading. It is what a table of contents entry and a cross reference point at. ,**An object's anchor is its segment name**, in the ordinary case, because a heading that opens a segment is arrived at by arriving at the segment. It differs only where a heading is not the opening of the segment it sits in: 69 of the 72 anchors in one reference document equal their segment and 3 do not. ,**A reader must arrive at the anchor whether or not the serialisation wrote it down**; see section 9. ## 8. Digests The model is checkable at four levels, and the chain is what lets an artefact be verified rather than trusted: ``` the markup files ──> a digest over every markup file of that language the serialisation ──> a digest of the .ssp a database was built from each object ──> a digest of the object's text each image ──> a digest of the image file, with its size in bytes and pixels ``` An object's digest is over its text alone, so a change to one paragraph moves one digest and no other. ## 9. Derived values, and why a reader must derive them Two of the values above are **derived**: the identifier from the OCN (section 3.1), and the anchor from the segment name (section 7.1). The model holds them; a serialisation need not, and the text form does not. ,**A serialisation may omit a value it can compute.** The `.ssp` writes a property only when it carries information, so it omits `.identifier` where the identifier is the OCN and `.anchor` where the anchor is the segment name. The `.ocda.db` is columns rather than lines and writes both unconditionally. Both are conforming. What is **not** conforming is a reader that takes the text form literally and hands on a model with those fields empty, because that model is not the same model the database gives, and a writer over it will produce different output from the two serialisations of one document. The size of the difference is easy to underestimate. In `free_culture.lawrence_lessig` the `.ssp` omits **1,314 identifiers and 69 anchors** that the `.ocda.db` states, which is nearly every object in the document. So, on loading: | field | when absent | take | |------------|--------------------------|------------------------------------------------------------------| | identifier | OCN is non-zero | the OCN as text | | identifier | OCN is `0` | whatever the serialisation gave; there is nothing to derive from | | anchor | the object has a segment | the segment name | ,**This is a property of the model, not of either format**, which is why it is stated here and not only in the two format specifications. It is stated here *because* two readers written independently from these documents disagreed for want of it: the rule was implied by the `.ssp`'s emission conditions and nowhere asserted. It is also what **spine's own reader** does, and always did, which is why the `.ocda.db` carries the derived values: the database is written from the `.ssp` read straight back, and the reading fills both in. So this section documents existing behaviour rather than prescribing new. It should have been written down first. ## 10. What the model does not hold - **The markup.** The model is what the markup produced, not the markup itself. An artefact that carries both is the `.ocda.db`, which carries the markup alongside the model rather than within it. - **The images.** The model *names* them, with digest, byte size and pixel size, and they must be found beside the file for a document to be complete. - **Any output format.** No HTML, no page numbers, no typesetting. #+END_SRC ** ssp-format-specification.md #+HEADER: :tangle ../ssp-format-specification.md #+BEGIN_SRC md # The `.ssp` format ,**The object-centric document abstraction, written as text. Format version 2.1**, as spine emits and reads it after 0.25.1. This document specifies *one serialisation*. The model it serialises is specified in [`ocda-model.md`](ocda-model.md), which the sqlite serialisation shares; read that first if the question is what an object ,*is* rather than how it is written down. ## The name `.ssp` belongs to the family of SiSU markup extensions, `.sst` (sisu text), `.ssm` (sisu master) and `.ssi` (sisu insert), and a short extension was wanted that sat with them. It stands loosely for **sisu peg**, or **sisu parsable**. `.ssa`, for sisu abstraction, was thought to be taken by other projects. So the extension names the *file family and its parsability*, where `.ocda.db` names the *model and its container*. The two are not parallel in construction and are not meant to be. ## Status This specification is derived from the emitter (`sisudoc/ocda/abstraction/ssp.d` in sisudoc-spine, whose `sspObjectRecord` is "the single definition of what an object looks like in a `.ssp` file"), and every claim in it is checked against the project's 36 document reference set. Where the two could differ, the emitter is normative and this document is wrong. ## 1. What a `.ssp` is A `.ssp` file is the serialisation, as text, of one language of one document after spine has parsed it: the object model, not the markup and not any output format. Spine's pipeline is **markup -> abstraction -> output**. A `.ssp` is the middle stage written down. Since spine 0.25.0 it is also read back, so a document may enter the pipeline at the abstraction stage instead of being parsed, and the output writers cannot tell which happened. The format is: - **line oriented.** Every construct is a whole line. No line continuations. - **order significant.** Sections appear in a fixed order, objects in document order, and an object's properties in a fixed order (section 6.3). - **UTF-8**, without a byte order mark. - **lossless for the abstraction**, not for the markup: a `.ssp` read back and re-emitted is byte identical, which the project's test suite asserts over the whole reference set. ### 1.1 What it is not It does not carry the markup it was made from, and it does not carry its images: it *names* them, with their digests and pixel sizes, and they must be found beside it (`../image/` within a pod) for the document to be complete. An artefact that carries both is the `.ocda.db`, which is built by reading a `.ssp` back, so it can hold nothing this format cannot express. ## 2. Version and compatibility The version is declared **in band**, on the first line of every file: ``` % SiSU Document Abstraction v2.1 ``` and by the sibling sqlite serialisation in its `schema.version` metadata row, so that one bump moves both. ### 2.1 The rule - The **minor** part is raised when a property is **added**. A reader of the newer format still reads the older one. - The **major** part is raised when a property **changes meaning or goes away**. - A reader **checks the major part and refuses what it does not know**, rather than loading an unfamiliar artefact half populated in silence. ### 2.2 History | version | date | what moved | |---------|------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 2.1 | 2026-09-29 | `.ocn_range`, the first and last ocn of what a container holds. A poem is a container of verses and is cited as their range: it now comes before its first verse, carries ocn `0` and no identifier, and states the range, where it had come after its last verse carrying that verse's ocn and identifier, so that one number named two objects. A minor bump: the property is an addition, the poem was never rendered and has no text, the verses are unchanged, and a 2.0 reader reads a 2.1 file. The same run gathers a note in a poem's last verse into the endnotes, which it had not been: one more endnote in the reference collection. And the editor's notes, `~[* ]~` and `~[+ ]~`, which spine had not recognised and passed through as text, are notes: a third form of note mark, a symbol and a number (`*1`, `+1`), each series numbered through the document; 135 more endnotes, 2,294 in all (section 3.2, "A note begins with its mark"). | | 2.0 | 2026-09-18 | `source.digest` changes meaning: it covers every markup file the document is made of, where it had been the digest of the master file alone. A change of meaning, so the major part. | | 1.1 | 2026-09-10 | The `@source` block; the document metadata the output writers read, carried in full (14 `@meta` and 3 `@make` properties added); `.segment_lv4_is`. | | 0.1 | | The first form. | ## 3. Lexical structure Six line types, distinguished by their first character. Nothing else is a valid line. | line | form | meaning | |--------------------|-----------------|-----------------------------------------------| | comment | `% ...` | file header only (section 4) | | block open | `@name {` | opens a preamble block or a section | | block close | `}` | closes it | | key | ` key: value` | inside a preamble block, two spaces of indent | | object declaration | `[ocn] kind...` | opens an object record (section 6.1) | | property | `.name: value` | a property of the object just declared | | text | `\| text` | a line of the object's text | | blank | | separates records; carries no meaning | A value runs to the **end of the line**. It is not quoted, and it is not trimmed beyond the single space after the colon. A value may therefore contain `:`, spaces, and any UTF-8 character except a line break. ### 3.1 Escaping Only one value is escaped, because only one value is a **joined list**: `.heading_ancestors_text` joins eight slots with `|`. A slot that itself contains `|`, a backslash, or a line break would otherwise be unreadable. | in the value | written as | |-----------------|------------| | `\` | `\\` | | `\|` | `\\|` | | newline | `\n` | | carriage return | `\r` | Unescaping reads left to right: a backslash escapes the next character, `\n` and `\r` become the control characters, any other escaped character stands for itself, and an unescaped `|` ends the slot. ,**No other value is escaped**, and text lines are not escaped at all: a text line is the object's text verbatim after the two-character `| ` prefix. ### 3.2 Object text is internal markup The text on `|` lines is spine's **internal** representation, neither the author's markup nor any output. It carries the markers spine uses to hold inline structure: The full set is defined in spine's `InlineMarkup` (`ocda/meta/defaults.d`). Counts are occurrences in object text across the 36 document reference collection, which distinguishes what a reader will certainly meet from what it may not. | marker | means | count | where | |-------------|---------------------------------------------|--------|----------------------| | `░` | non-breaking space | 23,048 | | | `┥` `┝` | link text, open and close | 20,290 | | | `┤` `├` | link target, open and close | 20,290 | | | `⑆` | a page number of the source edition follows | 19,099 | | | `┨` `┣` | that page number, open and close | 19,099 | | | `■` | monospace | 10,688 | | | `【` `】` | an inline note, its mark first (below) | 2,294 | | | `¤` | an internal site link | 2,157 | | | `┊` | table cell separator | 695 | `table` objects only | | `┙` | line break, inline | 184 | | | `☼` | an image, in a link's text slot | 63 | | | `┚` | line break, spaced | 50 | `group` objects only | | `〖` `〗` | a starred note | 0 | | | `‖` | a citation | 0 | | | `┘` | line break, plain | 0 | | | `┼` `┿` `╂` | page break: line, page, new page | 0 | | | `┏` `┆` | table open, table paragraph | 0 | | | `␣` | separator | 0 | | ,**A count of 0 means the reference collection does not exercise it, not that it cannot appear.** All of these are defined by the emitter and a reader must be prepared for them. ,**A note begins with its mark**, then whitespace, then the note: `【12 text】`. The mark is one of three forms, and a reader must take all three: | mark | what | example | |-------------------------|-------------------------------------------------------------------------------------------------------------------------------|----------------------------| | digits | the author's notes, numbered through the document | `【12 ...】` | | `*` or `+`, repeated | a special note, marked by the symbols alone | `【** ...】` | | `*` or `+`, then digits | an editor's note (2.1): the markup's `~[* ]~` and `~[+ ]~`, each series numbered through the document apart from the author's | `【*3 ...】`, `【+1 ...】` | So the pattern is `【(\d+|[*+]+[0-9]*)\s`. Before 2.1 the third form did not occur: spine did not recognise the editor's-note markup, and it came through as text, `~[* ...]~` and all. Each note of every form is gathered into the `endnotes` section, where its `.anchor_tag` is `note_` and the mark (`note_12`, `note_*3`). Two traps in this table, both found by counting rather than by reading: ,**`┚` is defined twice**, as the spaced line break and as the table close. In object text it occurs only in `group` objects, never in a `table`, so the object's kind resolves it. A reader should decide by kind rather than by character. ,**`「` and `」` are not markers here.** They are defined in `InlineMarkup` as the wrapper spine's ODT writer puts around an object number, but that happens at output, after the abstraction. Their 67 occurrences in the reference collection are **Japanese corner brackets in the document's own text**, being quotation marks in Japanese. In a `.ssp` they are always content. Some markers are the ASCII characters `*`, `!`, `/`, `_`, `^` and `,` for emphasis, bold, italic, underscore, superscript and subscript. These are ,**not distinguishable from ordinary text by inspection**, and a reader that needs them must follow spine's rules for where they apply rather than searching for the characters. A reader that renders text must decode all of this; a reader that only searches or counts need not, and one that wants plain text may strip the non-ASCII markers and take the ASCII ones as text. They are stable and they are part of the format. ## 4. File shape In order, and this order is fixed: ``` % SiSU Document Abstraction v2.1 % Source: @source { ... } @meta { ... } @make { ... } @doc_has { ... } @head { ... } @toc { ... } @body { ... } @endnotes { ... } @glossary { ... } @bibliography { ... } @bookindex { ... } @blurb { ... } @tail { ... } ``` The two `%` lines are the only comments in the file and appear only here. The four preamble blocks are always present and always in that order. The nine section blocks appear in that order. **A section is omitted entirely when the document has no objects for it**, though in practice that is rare: all 36 documents of the reference collection carry all nine, several of them holding a single generated object. ## 5. The preamble blocks Inside a block, every line is ` key: value` with exactly two spaces of indent. Keys are unique within a block. **A key whose value is empty is omitted**, so the set of keys present varies by document; the exceptions are noted below. ### 5.1 `@source` - what this abstraction was made from | key | always? | meaning | |-------------|----------------------------|-----------------------------------------------------------------------------| | `language` | yes | the language of *this* file | | `languages` | when the pod declares them | every language of the document, space separated | | `digest` | yes | sha256 over **every markup file the document is made of**, in a fixed order | The source *filename* is not repeated here; it is on the `% Source:` line. `digest` is what changed in 2.0. It had been the digest of the master file alone, which for a single file document was right and for a multi-file one was close to meaningless: any chapter could be edited without the digest moving. ### 5.2 `@meta` - the document's own metadata Dublin-Core shaped, `group.name`. **43 keys are defined**; a given document carries only those it has. The 28 the reference collection exercises are marked; the remainder are defined and simply unused by these documents. | group | keys | |---------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | `title.` | `main`*, `subtitle`*, `full`*, `language`*, `edition`* | | `creator.` | `author`*, `author_surname`*, `author_surname_fn`*, `author_email`*, `illustrator`*, `translator`* | | `date.` | `published`*, `created`*, `issued`*, `available`*, `modified`*, `valid`*, `added_to_site`* | | `rights.` | `copyright`*, `license`*, `copyright_text`, `copyright_translation`, `copyright_illustrations`, `copyright_photographs`, `copyright_cover`, `copyright_audio`, `copyright_video` | | `classify.` | `topic_register`*, `subject`, `keywords`, `loc`*, `dewey`* | | `identifier.` | `isbn`, `oclc` | | `notes.` | `summary`*, `abstract`*, `description`* | | `original.` | `source`*, `title`, `language`, `language_char` | | `language.` | `document_char`*, `document` | `classify.topic_register` is structured within its value: topics separated by `;`, levels within a topic by `:`, and alternatives at a level by `|`. ### 5.3 `@make` - how the document is to be built | key | value | meaning | |-------------------------|--------------------------|--------------------------------------------------------------------------------------------------------| | `doc_type` | `article`, `book` | what kind of document this is | | `auto_num_top_at_level` | a level, e.g. `1` | the heading level automatic numbering starts at, **as written in the markup** | | `auto_num_top_lv` | that level's index, 0..8 | the same level as a number (section 4 of [`ocda-model.md`](ocda-model.md)); `9` where numbering is off | | `auto_num_depth` | e.g. `2` | how many levels down numbering continues | | `breaks` | see below | where the output starts a new page, and where it breaks one | | `footer` | slots joined with `\|` | the footer, each slot carrying link markup | | `home_button_text` | link markup | the home link, as internal markup | `auto_num_top_at_level` and `auto_num_top_lv` are **the same value twice**, as the markup wrote it and as an index: `1` and `4`, since the body levels begin at index 4. `breaks` is a small language of its own, `new=` for a new page and `break=` for a page break, each taking a comma separated list of levels, the two separated by `;`: ``` breaks: new=:A,:B; break=:C,1 breaks: new=:B; break=1 breaks: break=1 ``` The level names here are written as they are in the markup, so with or without the leading colon as the document wrote them; a reader should not assume either form. ### 5.4 `@doc_has` - what the document contains, by count Ten keys, **all ten always emitted**, zero included. This block is a summary a reader can use without walking the objects. ``` inline_links inline_notes_reg inline_notes_star tables codeblocks images poems groups blocks quotes ``` ## 6. Objects The nine section blocks hold object records, in document order. An object record is an object declaration line, then its properties, then its text lines, then a blank line. ### 6.1 The declaration line ``` [] [] heading : [] heading : ``` `` is the **object citation number** (see [`ocda-model.md`](ocda-model.md) section 3). `[0]` means the object has none, is shared by many objects, and must never be used as a key; key on position within a section instead. `` is one of sixteen: ``` para heading toc bookindex endnote code bibliography comment verse glossary quote poem block table blurb group ``` A **heading** is the one kind whose declaration line carries more: its markup level after a colon (`:A` `:B` `:C` `:D` for the title levels, `:1`..`:4` for the body levels; `:D` is defined and unexercised by the reference collection), and then its identifier, when it has one that is not simply the ocn. Every other kind carries an identifier, on the same condition, as the `.identifier` property. > **A reader must derive the identifier where it is absent, and the anchor > likewise.** "when it has one that is not simply the ocn" means the > identifier *is* the ocn in every other case, and a reader that leaves the > field empty gives a different model from the one the `.ocda.db` gives, on > nearly every object. Section 9 of [`ocda-model.md`](ocda-model.md) states > both derivations and why they belong to the model rather than to either > format. Two readers written independently from these documents disagreed > for want of it. ### 6.2 Text lines ``` | ``` One line per line of the object's text, in order, each prefixed with `| ` (pipe, space). An object with no text has no text lines. The text is internal markup (section 3.2) and is not escaped. ### 6.3 Properties ,**Properties appear in a fixed order**, the order of the table below, and a reader may rely on it: checked over the reference set, no object among 56,882 emits a property out of this order. ,**A property is omitted when it holds its default**, so absence is meaningful and is defined per property in the table. Four properties repeat, appearing once per value rather than carrying a list: `.image`, `.stow_link`, `.lev4_subtoc` and `.anchor_tag`. Two consequences of the order, which a validator needs stated and which hold over the whole reference collection: - **a repeated property's occurrences are consecutive.** They share one position in the order, and nothing is written between them. Checked: no repeat in the collection is separated by another property. - **no property follows the text.** The record is declaration, then properties, then text lines, and once a `|` line has been written no further property appears. Checked: 0 occurrences in the collection. In the table, "emitted when" is the condition under which the property appears at all. Counts are occurrences across the 36 document reference set, as a guide to which parts of the format are load bearing. | # | property | value | emitted when | count | |----|---------------------------|-------------------------------|----------------------------------------------------------------------------------------------------|--------| | 1 | `.identifier` | string | a non-heading has an identifier that is not its ocn (**otherwise it is the ocn**: model section 9) | 353 | | 2 | `.part` | string | the object belongs to a part (`frontmatter`, `body`, ...) | 57,018 | | 3 | `.section` | string | the object's own section differs from the block it sits in | 605 | | 4 | `.parent` | int | **every object except** the document root and `toc` entries | 53,767 | | 5 | `.last_descendant` | int | non-zero: the last ocn under this heading | 3,202 | | 6 | `.ocn_range` | 2 ints | the object is a container: the first and last ocn of what it holds (2.1; a poem, model section 5) | 48 | | 7 | `.children` | ints | the heading has child headings | 827 | | 8 | `.ancestors` | 8 ints | any is non-zero | 53,731 | | 9 | `.ancestors_collapsed` | 8 ints | any is non-zero | 53,731 | | 10 | `.dom_status` | 8 ints | any is non-zero | 3,319 | | 11 | `.dom_status_collapsed` | 8 ints | any is non-zero | 3,319 | | 12 | `.heading_lev_collapsed` | int | it is below 9 | 3,503 | | 13 | `.parent_lev` | int | non-zero | 52,984 | | 14 | `.dummy` | `true` | the heading was introduced to segment an output | 183 | | 15 | `.ocn_off` | `true` | the object has no citation number | 7,543 | | 16 | `.is_of_type` | string | always, in practice | 57,018 | | 17 | `.attrib` | string | set | 1 | | 18 | `.meta_lang` | string | the object carries its own language | 0 | | 19 | `.meta_syntax` | string | the object carries its own markup syntax | 0 | | 20 | `.sha256` | hex | the object's text digest is non-zero | 51,253 | | 21 | `.image` | record | once **per image**, in order of appearance | 63 | | 22 | `.indent` | 2 ints | either indent is non-zero: `base hang` | 7,583 | | 23 | `.bullet` | `true` | the object is a bullet item | 1,903 | | 24 | `.lang` | string | the text attributes carry a language | 0 | | 25 | `.has` | flags | any is set | 11,384 | | 26 | `.table_cols` | int | the object is a table with columns | 32 | | 27 | `.table_widths` | ints | the table has column widths | 32 | | 28 | `.table_aligns` | letters | the table has alignments (`l` `c` `r`) | 32 | | 29 | `.table_header` | `true` | the table has a header row | 29 | | 30 | `.code_linenumbers` | `true` | the code block is numbered | 1 | | 31 | `.stow_link` | url | once **per extracted link** | 1,014 | | 32 | `.segment` | string | the heading opens a segment | 3,355 | | 33 | `.anchor` | string | the anchor differs from the segment name (**otherwise it is the segment name**: model section 9) | 68 | | 34 | `.segment_prev` | string | there is a previous segment | 1,022 | | 35 | `.segment_next` | string | there is a next segment | 1,022 | | 36 | `.heading_lev_anchor` | string | the heading has a level anchor tag | 1,854 | | 37 | `.segment_epub` | string | the heading opens an epub segment | 3,355 | | 38 | `.segment_html_is` | string | the object is inside an html segment | 47,072 | | 39 | `.segment_epub_is` | string | the object is inside an epub segment | 47,072 | | 40 | `.segment_lv4_is` | string | **and** it differs from `.segment_html_is` | 342 | | 41 | `.heading_ancestors_text` | 8 slots, `\|` joined, escaped | any slot is non-empty | 3,319 | | 42 | `.lev4_subtoc` | string | once **per sub-table-of-contents entry** | 1,887 | | 43 | `.anchor_tag` | string | once **per anchor tag** | 8,019 | Three of these need a word here, the first two because the model behind them is easy to misread (see [`ocda-model.md`](ocda-model.md) sections 5 and 7), the third because it is a property of this table rather than of the format. ,**`.parent: 0` is emitted, not omitted.** A parent of 0 is a real, non-substantive parent and not a missing value, so the property is written out rather than defaulted away: omitting it would make the two cases indistinguishable in the file. A reader walking upward stops at 0. ,**`.segment` and `.segment_html_is` are different questions**, and both are carried: what a heading opens, and what an object sits in. A reader wanting to know where a paragraph belongs needs the second. ,**Count 0 does not mean "not in the format".** `.meta_lang`, `.meta_syntax` and `.lang` are emitted by spine under the conditions given, and no document in the reference collection meets them. A reader must handle them; a test corpus does not exercise them. ### 6.4 The `.image` record ``` .image: sha256: bytes: px:x .image: missing:true ``` Filename first, as the object's handle on the image. The remaining fields are self describing `key:value` tokens, **so that further fields may be added without invalidating existing readers**; `px:` is omitted when the pixel size is unknown. `px:` is the file's own size, not the display size, which the object's text already carries. ### 6.5 The `.has` flags Space separated, in this order, each present only when set: ``` links notes_reg notes_star images images_no_dim ``` ## 7. Grammar ,**This grammar is complete for the file shape and deliberately incomplete for one thing: the order of properties.** A grammar that fixed the order would have to enumerate 42 optional terms in sequence, would be unreadable, and would need editing at every minor version, which is exactly when a property is added. The order is a guarantee of the format (section 6.3) and is stated there, where it can be read; a parser that wants to check it can do so against the table. So: **a file that this grammar accepts and section 6.3 rejects is a malformed file, not an ambiguity in the format.** The grammar is the shape; the table is the contract. ,**Checked**: every line of the 36 document reference collection matches one of the eight line shapes below, 721,363 lines with none left over. ```peg Document <- Header Source Meta Make DocHas Section* EOF Header <- "% SiSU Document Abstraction v" Version EOL "% Source: " Value EOL Blank Version <- Digit+ "." Digit+ Source <- "@source {" EOL Pair+ "}" EOL Blank Meta <- "@meta {" EOL Pair* "}" EOL Blank Make <- "@make {" EOL Pair* "}" EOL Blank DocHas <- "@doc_has {" EOL Pair+ "}" EOL Blank Pair <- " " Key ": " Value EOL Key <- [a-z_]+ ("." [a-z_]+)? Section <- "@" SectionName " {" EOL Blank Object* "}" EOL Blank SectionName <- "head" / "toc" / "body" / "endnotes" / "glossary" / "bibliography" / "bookindex" / "blurb" / "tail" Object <- Decl Property* TextLine* Blank Decl <- "[" Digit+ "] " (Heading / Kind) EOL Heading <- "heading :" Level (" " Identifier)? Level <- [A-D] / [1-4] Kind <- [a-z]+ Property <- "." Name ": " Value EOL Name <- [a-z_0-9]+ TextLine <- "| " Value EOL # a repeatable property appears once per value rather than carrying a list: # .image .stow_link .lev4_subtoc .anchor_tag # and .image has an inner shape of its own: Image <- ".image: " FileName (" " Field)* EOL Field <- Key ":" Value FileName <- (!" " !EOL .)+ Value <- (!EOL .)* Blank <- EOL EOL <- "\n" Digit <- [0-9] ``` ## 8. Conformance ### 8.1 A reader must - **Check the major version** on the first line and refuse a major it does not know. Loading an unknown major half populated, in silence, is the failure this rule exists to prevent. - **Tolerate unknown properties** within a known major version, since the minor part is raised precisely by adding them. Ignoring them is conforming behaviour. - **Treat `[0]` as not an identity.** Objects with no citation number are numerous and share the number. - **Stop at `.parent: 0`** when walking upward. ### 8.2 A reader should - **Report what it ignored**, at least on request. A reader that silently drops what it does not understand cannot tell its user whether it understood the document, and cannot be used to check this specification against an implementation. ### 8.3 A writer must - Emit the blocks and sections in the order of section 4, an object's properties in the order of section 6.3, and objects in document order. - Omit a property holding its default, per section 6.3. - Escape only what section 3.1 requires, and only where it requires it. ## 9. Relationship to the model, and to the sqlite serialisation The model this format writes down is specified in [`ocda-model.md`](ocda-model.md): objects, citation numbers, structure, sections, segments and digests. This document adds nothing to the model and subtracts nothing from it; it says how the model is written as text. The `.ocda.db` is the same abstraction as sqlite, and is **built by reading a `.ssp` back**: the document is emitted as `.ssp`, read straight in, and what comes back is written. So the database cannot carry an object property this format cannot express, and the two cannot drift. They share the version, which is why a bump moves both. The database is not a competitor to this format. It carries every language of a document in one file, and the markup, images and configuration as well, which this format does not. ## 10. Worked examples - `example-autonomous-contract.ssp` - one document, complete and unedited. - `example-wealth-of-networks.ssp` - an excerpt of a large one: verbatim lines with marked elisions, keeping the preamble, the `@head` section, and a run of body objects including one that embeds an image. Each carries, in its opening comments, the command that regenerates it. #+END_SRC ** ocda-db-format-specification.md #+HEADER: :tangle ../ocda-db-format-specification.md #+BEGIN_SRC md # The `.ocda.db` format ,**The object-centric document abstraction, written as sqlite. Format version 2.1**, as spine emits and reads it after 0.25.1. This document specifies *one serialisation*. The model it serialises is specified in [`ocda-model.md`](ocda-model.md), which the text serialisation shares; read that first if the question is what an object *is* rather than how it is stored. ## Status Derived from the writer, `sisudoc/outputs/io_out/sqlite_ocda_db.d`, and checked against the 27 published databases of the project's reference collection. Where the two could differ the writer is normative and this document is wrong. ## 1. What it is, and what makes it different One file per **document**, holding **every language of it**: a ten language document is one database with ten rows in `documents`, written one language at a time into the one file. Two things distinguish it from the `.ssp`: - **It carries its own source.** The markup, the images, `conf/document_make`, `pod.manifest` and the `tools/po4a` catalogues are all inside it, in the `files` table. A pod can be written back out of it, and spine does exactly that when asked for an artefact that needs markup. The `.ssp` names its images and carries none of them. - **It is queryable without being parsed.** A question about the document is a SELECT, answered from the file with no spine and no reader. ### 1.1 It is built from the `.ssp` The document is emitted as `.ssp`, **read straight back**, and what comes back is what is written. So **the database cannot carry an object property the `.ssp` cannot express**, and the two cannot drift. That is also why they share one version number: a bump moves both. The carried *files* are the deliberate exception. They come from the pod rather than from the abstraction, each with its digest. ## 2. Version Declared in band, in the `metadata` table: ```sql SELECT key, value FROM metadata WHERE doc_id IS NULL; -- build.spine_version | 0.25.0 -- schema.name | sisu-abstraction-db -- schema.version | 2.1 ``` The rule is the model's, and identical to the `.ssp`: the minor part is raised when a property is added, the major when one changes meaning or goes away, and **a reader checks the major part and refuses what it does not know**. See [`ssp-format-specification.md`](ssp-format-specification.md) section 2 for the history, which is shared. ## 3. The schema Eight tables, one virtual table, three views. ``` documents id, lang, source_filename, ssp_digest UNIQUE(lang) │ doc_id ├─ metadata doc_id, key, value UNIQUE(doc_id, key) │ └─ objects id, doc_id, section, seq, ocn, ... 50 columns UNIQUE(doc_id, section, seq) ├─ object_images object_id, seq, name, bytes, sha256, │ width, height, missing ├─ object_links object_id, seq, url ├─ object_anchors object_id, seq, anchor └─ object_subtoc object_id, seq, entry files id, role, name, bytes, sha256, width, height, compression, data UNIQUE(role, name) objects_fts fts5 over objects.text, external content over objects.id ``` ### 3.1 `doc_id` is the language key `documents.id`, referred to as `doc_id` everywhere else, is **what tells one language's rows from another's**. It is the key to reach for in every query. `objects.lang` and `objects.meta_lang` are **properties of the object** and are not it. They are almost always empty (section 6). ### 3.2 The file-level metadata rows `metadata` is keyed `UNIQUE(doc_id, key)` and not on key alone, because every language has a title and a creator and a key-only key would refuse the second. Three rows describe **the file** rather than a document in it, and are written with `doc_id` null: `schema.name`, `schema.version` and `build.spine_version`. In SQL no null equals another null, so `UNIQUE(doc_id, key)` does not constrain them, and a partial unique index over `key` alone `WHERE doc_id IS NULL` is what makes them singular. #### 3.2.1 The key prefix carries what the `.ssp` puts in blocks The `.ssp` has four preamble blocks, `@source`, `@meta`, `@make` and `@doc_has`. The database has **one** `metadata` table, and tells them apart by a **prefix on the key**: | `.ssp` block | database key | |--------------|-----------------------------------------------------------| | `@meta` | the key as it stands: `title.main`, `creator.author`, ... | | `@make` | `make.` + the key: `make.footer`, `make.doc_type` | | `@doc_has` | `doc_has.` + the key: `doc_has.images`, `doc_has.tables` | | `@source` | `source.` + the key: `source.language`, `source.digest` | ,**A reader must strip the prefix and restore the grouping**, or a writer over it has to ask `metadata["make.footer"]` from a database and `make["footer"]` from a `.ssp`, which is exactly the difference the two serialisations exist not to make. `source.` also carries two keys with no `@source` counterpart, `source.filename` and `source.ssp_digest`; the `.ssp` puts the first in its `% Source:` comment and cannot hold the second, which is a digest *of* it. Values may carry leading or trailing whitespace here where the `.ssp` cannot represent it, its `key: value` line being unable to distinguish a value that begins with a space from the separator. **Strip on load**, as a reader of the text form necessarily does. ### 3.3 Fixed-width arrays are JSON `children`, `ancestors`, `ancestors_collapsed`, `dom_status`, `dom_status_collapsed`, `heading_ancestors_text`, `table_widths` and `table_aligns` are stored as **JSON arrays**, with validity checks, so that `json_extract`, `json_each` and `json_array_length` reach inside them. > **Parse them as JSON.** A scan that strips `[`, `]`, `,` and `"` looks > adequate and is wrong on the first escape: the encoder writes `/` as > `\/`, so `FREE/OPEN-SOURCE SOFTWARE` comes back as > `FREE\/OPEN-SOURCE SOFTWARE`. All 36 document-languages of the reference > collection contain at least one, and the defect is invisible to any check > that does not compare the field itself. > > Note also that `table_widths` is an array of **integers except where the > columns divide evenly**, when it is `[33.3333,33.3333,33.3333]`. A decoder > that accepts strings and integers and then falls back to a default turns > a value it could not read into one that looks read. The **element types**, so that a decoder can be written to accept exactly what occurs: | column | elements | |--------------------------------------------------------------------------------------|-------------------------------------------------------------------| | `children`, `ancestors`, `ancestors_collapsed`, `dom_status`, `dom_status_collapsed` | integers | | `table_widths` | integers, or floats where the columns divide evenly | | `table_aligns` | strings, one of `l` `c` `r` | | `heading_ancestors_text` | strings, eight of them, empty where that level is not an ancestor | The `.ssp` writes the first three groups **space separated** and `heading_ancestors_text` **`|` separated**. The empty slots of the last are significant: they say which levels a heading sits under, so the separators carry as much as the text does. The open-ended lists are **rows in the four child tables** instead of columns: images, links, anchor tags and sub-table-of-contents entries. Each child table carries `seq`, which preserves the order they appeared in. ### 3.4 Indexes `objects(doc_id, section)`, `objects(ocn)`, `objects(parent_ocn)`, `objects(is_a)`, `objects(heading_level)` where it is a heading, and `files(role)`, `object_images(name)`. ## 4. The relationship to the `.ssp`, column by column ,**Nothing is lost in either direction.** Every `.ssp` property is a column or a child table; every column is a property, a value from the object declaration line, or a database key. Where the names differ they differ for a reason, the `.ssp` preferring the short form and the database the explicit one: | `.ssp` property | database | |--------------------|-----------------------| | `.parent` | `parent_ocn` | | `.last_descendant` | `last_descendant_ocn` | | `.dummy` | `dummy_heading` | | `.ocn_off` | `object_number_off` | | `.part` | `is_of_part` | | `.section` | `is_of_section` | Three are **expanded**, the `.ssp` folding what the database spreads: | `.ssp` property | database | |---------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------| | `.indent: ` | `indent_base`, `indent_hang` | | `.ocn_range: ` | `ocn_range_first`, `ocn_range_last` (2.1: `0` where the object is not a container; a 2.0 database has neither column, and reads as no range) | | `.has: links notes_reg notes_star images images_no_dim` | `has_links`, `has_notes_reg`, `has_notes_star`, `has_images`, `has_images_no_dim` | The four **repeatable** properties become the four child tables: | `.ssp` property | database | |-----------------|------------------| | `.image` | `object_images` | | `.stow_link` | `object_links` | | `.anchor_tag` | `object_anchors` | | `.lev4_subtoc` | `object_subtoc` | And four columns come from outside the property list: `ocn`, `is_a` and `heading_level` from the object declaration line, and `text` from its `|` lines. `id`, `doc_id` and `seq` are the database's own keys; `seq` is the object's position within its section, and `(doc_id, section, seq)` is what identifies an object when `ocn` will not: `0`, or, before 2.1, a poem sharing its last verse's number (see [`ocda-model.md`](ocda-model.md) section 3). ### 4.1 Where the two are not symmetric "Nothing is lost in either direction" is true of the model, and there are three places where the *files* nonetheless differ. A reader must know all three, because each of them made two independently written readers disagree. ,**1. The database states what the `.ssp` derives.** `identifier` and `anchor` are written on every row. The `.ssp` omits each where it can be worked out, the identifier being the ocn and the anchor the segment name (sections 6.1 and 6.3 of [`ssp-format-specification.md`](ssp-format-specification.md), section 9 of [`ocda-model.md`](ocda-model.md)). In `free_culture.lawrence_lessig` that is ,**1,314 identifiers and 69 anchors** stated here and absent there. Neither file is the poorer; a reader that does not derive is. ,**2. `translation.ocn_aligned` is here and cannot be in a `.ssp`.** It says whether the languages of a document share one numbering, and a `.ssp` is one language. It is therefore a fact about **the file**, not about the document in it, and it is the one such fact either serialisation carries. A reader should **keep it out of the document model**, for the same reason the `schema.*` and `build.*` rows are kept out: a writer over the model must not be able to tell which serialisation it was handed, and a field that exists only when the input was a database is exactly that tell. The model has no place for collection-wide facts yet; when it grows one, this belongs there. ,**3. Whitespace and grouping**, as section 3.2.1 gives them. ## 5. The `files` table: what it carries, and how | role | named by | `compression` | `data` holds | |------------|----------------------------|---------------|------------------| | `image` | its bare filename | `none` | the file as read | | `source` | `media/text//` | `zstd` | a zstd frame | | `conf` | `conf/document_make` | `zstd` | a zstd frame | | `manifest` | `pod.manifest` | `zstd` | a zstd frame | | `tools` | `tools/po4a/...` | `zstd` | a zstd frame | The `compression` column carries the literal `none` or `zstd`, so a reader need not infer it from the role. Three rules make the file predictable: 1. **Compression is decided by role, never by size.** The same kind of file always stores the same way, so nobody need measure a document to know what its database will look like. A very small file can come out a few bytes larger inside a zstd frame; that is accepted rather than fixed with a threshold, which would make two documents of the same kind store differently for no reason a reader could see. 2. **Images are not compressed.** png and jpeg are compressed already; zstd gains nothing and costs a decompression on every read. 3. **`objects.text` is never compressed.** The full text index reads that column directly, and compressing it would stop search working completely rather than subtly. ,**`bytes` and `sha256` are always those of the original file**, whatever `compression` says. So a digest check, a `digests.txt` line and a rebuilt `source.digest` all keep working, and a reader that does not decompress can still say what it is looking at. ## 6. Querying it `sqlite3` is all that is needed. The recipes below were run against the published databases. ### 6.1 What is this file ```sql SELECT key, value FROM metadata WHERE doc_id IS NULL; SELECT id, lang, source_filename FROM documents ORDER BY id; SELECT role, count(*), sum(bytes) FROM files GROUP BY role; ``` ### 6.2 The views | view | gives | |------------------|---------------------------------------------------------------------------------------------| | `outline` | the headings: lang, section, seq, ocn, both levels, `last_descendant_ocn`, identifier, text | | `citable` | every object with an ocn, ordered by ocn, language named | | `document_files` | what the file carries: role, name, bytes, sha256, pixel size | ,**`outline` and `citable` name the language and order by it first.** A view over a file that may hold ten languages and does not say which would read as one document and be several. ,**A trap in `citable`:** it is `SELECT d.lang AS lang, o.*`, and `objects` has a `lang` column of its own, so sqlite renames the second one. The view has **`lang`, the document's language, and `lang:1`, the object's own**. Quote the second if you want it: `SELECT "lang:1" FROM citable`. ### 6.3 Structure A heading's subtree is a **range**, not a traversal, because every heading carries `last_descendant_ocn`: ```sql SELECT count(*) FROM objects o, (SELECT doc_id, ocn, last_descendant_ocn FROM objects WHERE is_a='heading' AND ocn=66) h WHERE o.doc_id = h.doc_id AND o.ocn BETWEEN h.ocn AND h.last_descendant_ocn; ``` And the path back up, on `parent_ocn`: ```sql WITH RECURSIVE up(ocn, parent_ocn, text) AS ( SELECT ocn, parent_ocn, text FROM objects WHERE doc_id=1 AND ocn=345 UNION ALL SELECT o.ocn, o.parent_ocn, o.text FROM objects o JOIN up ON o.ocn = up.parent_ocn AND o.doc_id=1 AND o.is_a='heading' WHERE up.parent_ocn > 0) SELECT ocn, substr(text,1,44) FROM up; ``` ,**Keep the `parent_ocn > 0` guard.** A parent of 0 is a real, non-substantive parent and not a missing value, and without the guard the recursion does not terminate. ### 6.4 Full text ```sql SELECT o.ocn, substr(o.text,1,44) FROM objects_fts f JOIN objects o ON o.id = f.rowid JOIN documents d ON d.id = o.doc_id WHERE f.text MATCH 'commons' AND d.lang='en' AND o.ocn > 0 ORDER BY rank LIMIT 3; ``` fts5 syntax is available in full: `MATCH 'peer AND production'`, `MATCH '"peer production"'`, `MATCH 'peer NEAR production'`. The index is **external content** over `objects.id`: the text is indexed, not stored twice. ### 6.5 Three things to know before trusting a query 1. **An ocn is not unique within a document**, and `0` is shared by many objects. Key on `objects.id`, or on `(doc_id, section, seq)`. 2. **`objects.text` is internal markup**, neither the author's markup nor any output. Fine for search and for reading structure, not for display without decoding. The markers are in [`ssp-format-specification.md`](ssp-format-specification.md) section 3.2. 3. **A multi-language file answers an unfiltered query with every language at once.** A single-language document has one `documents` row, so plain queries read naturally there; anything else wants `WHERE lang='..'` or a join to `documents`. ## 7. Digests, and verifying a file ``` the markup files ──> metadata source.digest sha256 over every markup file of that language the .ssp ──> documents.ssp_digest of the .ssp it was built from each object ──> objects.sha256 over the object's text each carried file──> files.sha256 over the original bytes the spine used ──> metadata build.spine_version ``` So `.sst -> .ssp -> .ocda.db` is checkable end to end, and `spine --ocda-verify=` closes the loop by writing the carried markup back out and re-parsing it: **does this spine still make this abstraction from the markup this file carries?** It writes nothing and the exit status is the answer. ,**Only this artefact can answer that from itself alone**, which is the strongest sense in which it is the one to work with. The database's own digest is taken once, when the last language has been written, and appears in `digests.txt`. The metadata page links the database and claims no hash for it, deliberately: the page is written per language while the database is added to per language, so a digest taken there would be of a file that no longer exists by the end of the run. ## 8. Conformance A reader **must** check `schema.version`'s major part and refuse a major it does not know; **must** treat `ocn` as not an identity; **must** stop at `parent_ocn = 0` when walking upward; and **should** say what it did not understand, for the reason given in [`ssp-format-specification.md`](ssp-format-specification.md) section 8.2. A writer **must** keep `bytes` and `sha256` those of the original file whatever `compression` says, and **must not** compress `objects.text`. #+END_SRC * COPYRIGHT & LICENSE ** notices *** project (project root) ./ #+HEADER: :tangle "../COPYRIGHT" #+HEADER: :noweb yes #+BEGIN_SRC org - Name: spine - SiSU Spine, Doc Reform <> <> <> <> #+END_SRC *** code org ./org - .org literate programming: - copyright asserted - omitted - license - dependency information #+HEADER: :tangle "../org/COPYRIGHT" #+HEADER: :noweb yes #+BEGIN_SRC text - Name: spine - SiSU Spine, Doc Reform <> <> <> <> #+END_SRC ** incorporate *** copyright #+NAME: sisudoc_spine_copyright #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:sisudoc_spine_copyright()>> #+END_SRC *** agpl3 #+NAME: sisudoc_spine_license_agpl3 #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:sisudoc_spine_license_agpl3()>> #+END_SRC *** spine_summary #+NAME: sisudoc_spine_summary #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:sisudoc_spine_summary()>> #+END_SRC *** markup_samples #+NAME: sisudoc_spine_markup_samples #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:sisudoc_spine_markup_samples()>> #+END_SRC * org includes ** project version #+NAME: spine_version #+HEADER: :noweb yes #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:spine_project_version()>> #+END_SRC ** year #+NAME: year #+HEADER: :noweb yes #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:year()>> #+END_SRC ** document header including copyright & license #+NAME: doc_header_including_copyright_and_license #+HEADER: :noweb yes #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:spine_doc_header_including_copyright_and_license()>> #+END_SRC * nixpkgs source #+NAME: nixpkgs_unstable #+HEADER: :noweb yes #+BEGIN_SRC emacs-lisp <<./sisudoc_spine_version_info_and_doc_header_including_copyright_and_license.org:nixpkgs_unstable()>> #+END_SRC * __END__