omni-viewer-core 0.18.0 → 0.20.0

package/README.md

@@ -13,11 +13,14 @@ models, viewers mount into a DOM element, and everything host-specific
 ## Supported formats
 
 - **Documents** — PDF, Word (DOCX and legacy DOC), HWP, PowerPoint (PPTX and
-  legacy PPT), Markdown, LaTeX (structure and math preview, not typesetting)
-- **Data & spreadsheets** — Excel, CSV/TSV, JSON, JSONL/NDJSON, YAML, TOML,
+  legacy PPT), Markdown, Jupyter Notebook (.ipynb), EPUB e-books (2 and 3),
+  LaTeX (structure and math preview, not typesetting)
+- **Data & spreadsheets** — Excel, CSV/TSV, JSON, JSONL/NDJSON, YAML, TOML, XML,
   Parquet, Avro, HDF5, MATLAB MAT, NumPy (NPY/NPZ), Safetensors, GGUF, ONNX, TFLite/LiteRT,
   Keras (.keras and legacy .h5), Core ML (.mlmodel and .mlpackage),
-  OpenVINO IR (.xml + .bin), PyTorch Export (.pt2), Protocol Buffers, ReqIF, SQLite
+  OpenVINO IR (.xml + .bin), PyTorch Export (.pt2), ExecuTorch (.pte), Protocol
+  Buffers, ReqIF, SQLite
+- **Web & network** — HAR (HTTP Archive) request logs
 - **Media & graphics** — audio (waveform/spectrogram), video, images,
   Photoshop PSD
 - **Engineering & automotive** — CAN DBC, AUTOSAR ARXML, ASAM A2L, Vector
@@ -299,6 +302,217 @@ through `probeContainer`, which recognizes the `archive_format` marker
 A plain `torch.save` checkpoint (`.pt` / `.pth`) is not an Export package and
 is not claimed.
 
+### ExecuTorch programs (.pte)
+
+A `.pte` is what `to_executorch().save()` writes: a FlatBuffer following
+ExecuTorch's `program.fbs` (file identifier `ET12`), optionally followed by data
+segments — constant tensors, delegate blobs, mutable initial state, and named
+blobs — that an extended header (`eh00`, inserted right after the identifier)
+locates. The parser reads the FlatBuffer directly with no schema compiler and no
+runtime dependency, and never decodes a segment or buffer payload:
+
+```ts
+import { mountPteViewer } from 'omni-viewer-core/viewers/pte';
+
+await mountPteViewer({ fileName: file.name, data: bytes }, container, ctx);
+```
+
+`parsePte` keeps every method (`ExecutionPlan`) with its values, instructions,
+operators, and delegates, and resolves where each tensor's bytes live: the
+deprecated inline `constant_buffer`, the constant segment (`constant_segment`
+offsets, with the alignment-padded slot size and absolute file offset), a
+mutable data segment for buffers with initial state, an external `.ptd` file
+(`ExtraTensorInfo.location = EXTERNAL`), a memory-planned arena slot
+(`AllocationDetails`), or nothing at all for runtime inputs. A kernel call's
+outputs come from the return value the emitter appends after the schema
+arguments (the out tensor, a TensorList of the out tensors, or a fresh scalar
+for a symbolic op), and only the out-argument positions in front of it are
+dropped from the inputs; a delegate call carries inputs then outputs with no marker,
+so those are split by data flow — the trailing arguments nothing defined yet.
+A view or alias the emitter gave its own EValue — a tensor nothing writes that
+shares a memory-planned slot with an earlier one — is resolved to the value it
+aliases when the file leaves no other reading, so a delegate fed through
+aliases keeps its inputs and the graph keeps its edges. Every EValue kind
+gets a preview, delegate compile specs are decoded when they are short text,
+and the leading instructions keep their emitter stack traces.
+
+The viewer draws each method's kernel, delegate, and move calls as a graph with
+inputs, consumed constants, and outputs, plus an inspector (arguments with their
+values, backend and compile specs for a delegate call, jump destination, stack
+trace), and searchable instruction, value, input/output, delegate, segment, and
+model-information panels; a multi-method program gets a method selector.
+Warnings cover backends the runtime must register, tensors stored in a separate
+`.ptd`, segments the file cannot locate or that run past its end, constant data
+the program never provides, references to segments the program never declares,
+and unrecognized scalar types, value kinds, or instruction kinds.
+
+Routing is by the `.pte` extension, validated against the `ET12` identifier at
+offset 4; an extensionless program is detected from the same bytes. A `.ptd`
+data file (`FT01`) is not a program and is not claimed.
+
+### Jupyter notebooks (.ipynb)
+
+View nbformat 4 notebooks with their code and saved results in file order.
+Markdown cells reuse the Markdown renderer, syntax highlighter, math renderer,
+and prose styles. Code cells show `In`/`Out` execution counts, stdout/stderr,
+HTML tables, PNG/JPEG/GIF/WebP/SVG images, Markdown, LaTeX, JSON, plain text,
+and error tracebacks. Raw cells display as text. Code is never executed.
+
+```ts
+import { mountNotebookViewer } from 'omni-viewer-core/viewers/notebook';
+import { loadNotebookViewerDeps } from 'omni-viewer-core/viewers/notebook/self-loading';
+
+const deps = await loadNotebookViewerDeps();
+const handle = await mountNotebookViewer(
+  { fileName: 'analysis.ipynb', data: notebookBytes }, container, hostContext, deps
+);
+// Release the viewer, listeners, and any host-resolved assets on close.
+handle.dispose();
+```
+
+The loader requires the existing optional peers `marked` and `dompurify`;
+`highlight.js` and `katex` enhance code and formulas when installed. Adapters
+can inject the same dependencies they use for Markdown. Load KaTeX CSS and
+fonts inside the viewer's shadow root when using KaTeX, as with Markdown.
+Scoped mounting uses `styles/notebook.css`.
+
+Search cells and results, hide all code or outputs, or expand individual cells.
+Persisted `source_hidden`, `outputs_hidden`, and `collapsed` hints set the
+initial expansion state. One static representation is selected per MIME bundle;
+interactive widget/JavaScript outputs fall back to saved plain text, or a
+clear unsupported-output notice. Markdown `attachment:` images work offline.
+Relative images use the optional `documentAssets` service, external links use
+`navigation`, and remote resources and active HTML are blocked.
+
+`parseNotebook` is also available from `parsers/notebook`. It returns a typed
+`ParseOutcome<NotebookDocument>` and uses the core JSON parser, retaining raw
+JSON output tokens. Defaults: 64 MiB input, 2,000 cells, 10,000 outputs, 4 MiB
+per preview, and 16 MiB cumulative previews. `limits.maxEntries` counts cells;
+`maxOutputs` and `maxTotalPreviewBytes` are independent limits. Truncation
+returns `partial` with diagnostics. Invalid JSON and unsupported major versions
+return typed failures. See [viewer details](examples/notebook/README.md) and the
+[sample notebook](examples/notebook/analysis.ipynb).
+
+### HAR network logs (.har)
+
+A `.har` file is the HTTP Archive a browser's network panel (or curl, Charles,
+Fiddler, Playwright) exports: one JSON `log` holding every request of a session
+with its headers, bodies, sizes, and phase timings. It is the artifact a web or
+API failure usually arrives as, so the viewer is built for triage rather than
+for reading JSON:
+
+```ts
+import { mountHarViewer } from 'omni-viewer-core/viewers/har';
+
+await mountHarViewer({ fileName: file.name, data: bytes }, container, ctx);
+```
+
+`parseHar` is a contract parser (`ParseOutcome<HarDocument>`) and reads the
+archive through the core JSON layer, so a truncated export still yields the
+requests it did hold. Each entry keeps its request and response headers, query
+string, cookies, request payload, and response body, and gains what the table
+needs: host, path, status class, resource type (from `_resourceType` when the
+writer supplied one, from the MIME type otherwise), transferred bytes
+(`_transferSize` when present), decoded resource size, and a phase breakdown
+with the TLS handshake taken out of `connect` so a waterfall cannot
+double-count it. Timestamps go through a core ISO-8601 parser rather than
+`new Date(string)`, and the timeline's origin is the earliest entry, so offsets
+and bar positions are the same on every platform. A base64 body is decoded when
+its MIME type is textual (the URL-safe alphabet included) and reported by size
+only when it is not; bodies are kept up to a preview limit, and entries up to
+`HAR_DEFAULT_MAX_ENTRIES` (a larger archive comes back `partial` with a
+diagnostic). A body cut by the preview limit and one the archive stored
+incompletely — several padded base64 chunks, or a tail outside the alphabet —
+are reported as two different things, and a size or start time the archive never
+gave is kept unknown rather than shown as a confident zero.
+
+The viewer shows summary figures (requests, elapsed, transferred, resources,
+domains, errors), a request table sortable by every column with a per-entry
+waterfall bar, and filters for method, status class, resource type, domain, and
+free text over URLs, status text, and headers — with an opt-in toggle that
+extends the search to bodies. Selecting a request opens headers (general,
+request, response), payload (query string and request body, form fields
+expanded), response (body pretty-printed through the JSON serializer when it
+really is JSON), cookies, and timings (drawn phases plus the values the archive
+reported, with `-1` shown as not applicable). Page load timings and archive
+metadata get their own tabs. Routing is by the `.har` extension; an
+extensionless archive is recognized by content — a `log` object holding an
+`entries` array — and claimed before the generic JSON viewer.
+
+### EPUB e-books (.epub)
+
+EPUB 2.0.1 and EPUB 3.x books open in a reflowable reading pane with the
+book's table of contents, chapter navigation, text scaling, and a book-info
+panel (metadata, cover, diagnostics). Sanitizing chapter markup is mandatory,
+so the viewer takes a DOMPurify factory — `loadEpubViewerDeps` imports the
+optional `dompurify` peer for you:
+
+```ts
+import { mountEpubViewer } from 'omni-viewer-core/viewers/epub';
+import { loadEpubViewerDeps } from 'omni-viewer-core/viewers/epub/self-loading';
+
+await mountEpubViewer({ fileName: file.name, data: bytes }, container,
+    { ...ctx, navigation }, await loadEpubViewerDeps());
+```
+
+`parseEpub` is a contract parser (`ParseOutcome<EpubDocument>`) that reads
+only the package files — `container.xml`, the OPF package document, the EPUB 3
+navigation document or the EPUB 2 NCX, and `encryption.xml` — so opening a
+large illustrated book costs its package files, not its images. EPUB 2
+`opf:` attributes and EPUB 3 `refines` metadata map onto one model (title and
+subtitles, creators with roles and sort names, identifiers with schemes,
+publication date, series, subjects, plain-text description). Chapters and
+images are inflated on demand through `openEpubArchive`, each read capped per
+member and against a cumulative budget, and a member that inflates past the
+size its directory record declares is rejected as corrupt. Inflation uses the
+platform's `DecompressionStream('deflate-raw')`; pass `inflate` to supply one
+on hosts without it.
+
+Chapter markup is parsed and sanitized in an inert document before it reaches
+the page: scripts, forms, frames, media, and **publisher CSS** (`<link>`,
+`<style>`, `style=""`) are dropped and the viewer's own reading typography
+applies; container images become blob URLs revoked on chapter change; remote
+images are blocked and counted. Internal links move through the spine,
+fragments scroll within the chapter, and external links go only through the
+host `navigation` service (allow-listed schemes). Next/previous walk the linear
+spine (`linear="no"` covers and notes stay reachable by link), arrow keys turn
+pages in the book's progression direction, and `Ctrl/Cmd` `+`/`-`/`0` scale
+text. A book whose every chapter is DRM-encrypted fails with
+`password-required` (font obfuscation is ignored); partly encrypted, missing,
+or fixed-layout content is reported as diagnostics. Routing is by the `.epub`
+extension; an extensionless book is recognized by `probeContainer` from its
+stored `mimetype` member.
+
+### XML (.xml and XML dialects)
+
+Configuration and integration documents (`pom.xml`, Android manifests,
+MSBuild projects, SOAP/WSDL, RSS/Atom, XLIFF, KML/GPX…) open as an
+element/attribute tree with a node inspector and the original source side by
+side:
+
+```ts
+import { mountXmlViewer } from 'omni-viewer-core/viewers/xml';
+
+const handle = await mountXmlViewer({ fileName: file.name, data: bytes }, container, { ...ctx, clipboard });
+```
+
+`parseXml` is a contract parser (`ParseOutcome<XmlDocument>`) with no DOM
+dependency. Every node keeps its source range, and namespaces are resolved.
+Well-formedness errors (mismatched or unclosed tags, duplicate attributes,
+undeclared prefixes…) are reported with line and column, then recovered, so a
+broken file still opens as a `partial` tree. External entities are never
+loaded. Internal DTD entities expand one level only, under a budget, which
+rules out XXE and billion-laughs expansion.
+
+Selecting a node highlights its range in the source. Clicking a tag or
+attribute in the source selects that node. Search covers names, values, or
+comments and PIs, with next/previous and a matches-only filter. Copy gives the
+XPath (positions appear only where names repeat), the value, or the node's XML.
+Both panes are virtualized, and long single-line sources are split into
+display segments. Defaults are 64 MiB input, 500,000 nodes, and depth 1,000.
+ARXML and ReqIF keep their dedicated viewers by extension. An `.xml` whose root
+is an OpenVINO IR goes to the OpenVINO viewer when a text sample is provided.
+
 ### Archive host integration
 
 Archive decompression remains adapter-owned: pass an `ArchiveDecoder` to

package/dist/i18n/catalog.en.js

@@ -4,6 +4,36 @@
 // this catalog at build time. Diagnostics (§3-① messageKey) and viewer UI
 // strings share this key space.
 export const CATALOG_EN = {
+    'notebook.title': 'Jupyter Notebook',
+    'notebook.summary': '{cells} / {total} cells · {outputs} outputs · {errors} errors',
+    'notebook.savedOutputs': 'Saved results · Code is displayed without execution.',
+    'notebook.search': 'Search cells and outputs',
+    'notebook.showCode': 'Show code',
+    'notebook.showOutputs': 'Show outputs',
+    'notebook.cells': 'Notebook cells',
+    'notebook.cell': 'Cell {index}',
+    'notebook.markdown': 'Markdown',
+    'notebook.code': 'Code',
+    'notebook.raw': 'Raw',
+    'notebook.unknown': 'Unknown cell',
+    'notebook.outputs': 'Outputs ({count})',
+    'notebook.noOutputs': 'No saved outputs.',
+    'notebook.empty': 'This notebook has no cells.',
+    'notebook.noMatches': 'No matching cells.',
+    'notebook.imageOutput': 'Saved image output',
+    'notebook.assetUnavailable': 'Image unavailable',
+    'notebook.unsupportedOutput': 'No supported static representation is available for this output.',
+    'notebook.truncatedOutput': 'This output exceeds the preview limit.',
+    'notebook.renderFailed': 'Preview failed. The source is shown below.',
+    'notebook.blockedResources': 'External resources were blocked.',
+    'diag.notebook.invalid': 'Invalid Jupyter Notebook document.',
+    'diag.notebook.version': 'Only nbformat version 4 notebooks are supported.',
+    'diag.notebook.malformed': 'Some malformed cells or outputs could not be read.',
+    'diag.notebook.unknownCell': 'An unknown cell type is shown as source.',
+    'diag.notebook.unknownOutput': 'An unknown output type could not be rendered.',
+    'diag.notebook.cellLimit': 'The cell preview limit was reached. Some cells are omitted.',
+    'diag.notebook.outputLimit': 'The output preview limit was reached. Some outputs are omitted.',
+    'diag.notebook.previewLimit': 'Some content exceeds the preview size budget and was truncated.',
     'audio.title': 'Audio', 'video.title': 'Video', 'dbc.title': 'CAN database',
     'audio.play': 'Play', 'audio.pause': 'Pause', 'audio.stop': 'Stop',
     'audio.volume': 'Volume',
@@ -215,6 +245,324 @@ export const CATALOG_EN = {
     'tflite.inspector.inputs': 'Inputs', 'tflite.inspector.outputs': 'Outputs',
     'tflite.inspector.intermediates': 'Intermediates', 'tflite.inspector.options': 'Options',
     'tflite.inspector.customOptions': 'Custom options', 'tflite.inspector.subgraphs': 'Referenced subgraphs',
+    'pte.title': 'ExecuTorch program', 'pte.search': 'Search instructions, values, delegates, and segments…',
+    'pte.copyJson': 'Copy JSON', 'pte.method': 'Method', 'pte.unnamed': 'unnamed',
+    'pte.graph': 'Graph', 'pte.instructions': 'Instructions', 'pte.values': 'Values',
+    'pte.io': 'Inputs & Outputs', 'pte.delegates': 'Delegates', 'pte.segments': 'Segments', 'pte.modelInfo': 'Model Info',
+    'pte.rows': '{shown} / {total} rows shown', 'pte.noMatches': 'No rows match the current search.',
+    'pte.selectNode': 'Select an instruction or value to inspect it.',
+    'pte.emptyGraph': 'This method has no kernel or delegate calls.',
+    'pte.graphLimited': 'Graph preview shows {shown} of {total} items.',
+    'pte.graphEdgesLimited': 'Only the first {count} connections are shown.',
+    'pte.moreItems': '{count} more…', 'pte.schema': 'schema v{version}',
+    'pte.scalar': 'scalar', 'pte.none': 'None', 'pte.yes': 'yes',
+    'pte.instructionDetail': '#{index} · chain {chain}', 'pte.jumpTo': 'jump to #{index}',
+    'pte.jumpEnd': 'jump past the end of this chain',
+    'pte.jumpInvalid': 'jump to {index}, outside this chain',
+    'pte.binaryBytes': '{bytes} bytes (binary)', 'pte.missingDelegate': 'Delegate {index} is not declared by this method.',
+    'pte.inRange': 'In file', 'pte.outOfRange': 'Past end of file',
+    'pte.unlocated': 'Unlocated (no header)',
+    'pte.copyFailed': 'Copy failed',
+    'pte.summary.methods': 'Methods', 'pte.summary.instructions': 'Instructions',
+    'pte.summary.operators': 'Operators', 'pte.summary.tensors': 'Tensors',
+    'pte.summary.delegates': 'Delegates', 'pte.summary.constants': 'Constant data',
+    'pte.warning.identifier': 'The file identifier is {identifier} rather than the expected ET12.',
+    'pte.warning.noMethods': 'The program contains no methods.',
+    'pte.warning.noInstructions': 'The program methods contain no instructions.',
+    'pte.warning.headerMissing': '{count} data segment(s) are declared but the file has no extended header to locate them.',
+    'pte.warning.segmentsOutOfRange': '{count} data segment(s) extend past the end of the file.',
+    'pte.warning.delegates': '{count} delegate call(s) need these backends registered in the runtime: {names}',
+    'pte.warning.externalTensors': '{count} tensor(s) are stored in a separate .ptd file that this viewer does not read.',
+    'pte.warning.missingData': '{count} tensor(s) reference constant data the program does not provide.',
+    'pte.warning.missingSegments': '{count} reference(s) name a data segment the program does not declare.',
+    'pte.warning.unknownScalarTypes': '{count} tensor(s) use a scalar type this viewer does not recognize.',
+    'pte.warning.unknownValueKinds': '{count} value(s) are of a kind this viewer does not recognize.',
+    'pte.warning.unknownInstructions': '{count} instruction(s) are of a kind this viewer does not recognize.',
+    'pte.column.index': '#', 'pte.column.name': 'Name', 'pte.column.chain': 'Chain', 'pte.column.kind': 'Kind',
+    'pte.column.target': 'Operator / Backend', 'pte.column.inputs': 'Inputs', 'pte.column.outputs': 'Outputs',
+    'pte.column.args': 'Args', 'pte.column.type': 'Type', 'pte.column.shape': 'Shape', 'pte.column.typeShape': 'Type / Shape',
+    'pte.column.elements': 'Elements', 'pte.column.dataSize': 'Data size', 'pte.column.storage': 'Storage',
+    'pte.column.memory': 'Memory plan', 'pte.column.valueIndex': 'Value', 'pte.column.backend': 'Backend',
+    'pte.column.location': 'Location', 'pte.column.dataIndex': 'Data index', 'pte.column.fileOffset': 'File offset',
+    'pte.column.compileSpecs': 'Compile specs', 'pte.column.offset': 'Offset', 'pte.column.status': 'Status', 'pte.column.usedBy': 'Used by',
+    'pte.kind.KernelCall': 'Kernel call', 'pte.kind.DelegateCall': 'Delegate call', 'pte.kind.MoveCall': 'Move',
+    'pte.kind.JumpFalseCall': 'Jump if false', 'pte.kind.FreeCall': 'Free', 'pte.kind.Unknown': 'Unknown',
+    'pte.kind.input': 'Input', 'pte.kind.output': 'Output',
+    'pte.storage.inline': 'Inline constant', 'pte.storage.segment': 'Constant segment', 'pte.storage.mutable': 'Mutable state',
+    'pte.storage.external': 'External file', 'pte.storage.planned': 'Memory planned', 'pte.storage.runtime': 'Runtime',
+    'pte.location.inline': 'Inline', 'pte.location.segment': 'Segment',
+    'pte.segment.constant': 'Constant tensors', 'pte.segment.mutable': 'Mutable state', 'pte.segment.delegate': 'Delegate blob',
+    'pte.segment.named': 'Named data', 'pte.segment.unused': 'Unreferenced',
+    'pte.info.identifier': 'File identifier', 'pte.info.schemaVersion': 'Schema version', 'pte.info.fileSize': 'File size',
+    'pte.info.extendedHeader': 'Extended header', 'pte.info.noHeader': 'none (no data segments)',
+    'pte.info.headerDetail': 'program {program} · segment base {base} · segment data {data}',
+    'pte.info.constants': 'Constant data', 'pte.info.inlineBuffers': '{count} inline buffer(s)',
+    'pte.info.delegateData': 'Delegate data', 'pte.info.backends': 'Backends', 'pte.info.methods': 'Methods',
+    'pte.info.namedData': 'Named data', 'pte.info.memoryPlan': 'Memory plan', 'pte.info.inputSpec': 'Input spec',
+    'pte.info.outputSpec': 'Output spec', 'pte.info.operator': 'Operator',
+    'pte.info.dimOrder': 'Dim order', 'pte.info.expectedSize': 'Expected size', 'pte.info.shapeDynamism': 'Shape dynamism',
+    'pte.info.segment': 'Segment', 'pte.info.dataIndex': 'Data index', 'pte.info.storageOffset': 'Storage offset',
+    'pte.info.requiresGrad': 'Requires grad', 'pte.info.device': 'Device',
+    'pte.info.aliasOf': 'Shares memory with',
+    'pte.inspector.inputs': 'Inputs', 'pte.inspector.outputs': 'Outputs', 'pte.inspector.arguments': 'Arguments',
+    'pte.inspector.delegate': 'Delegate', 'pte.inspector.destination': 'Destination', 'pte.inspector.stackTrace': 'Stack trace',
+    'pte.inspector.value': 'Value', 'pte.inspector.items': 'Items',
+    'xml.title': 'XML document',
+    'xml.eyebrow': 'XML {version}',
+    'xml.unreadable': 'This file could not be read as XML.',
+    'xml.search': 'Search names, values, comments…',
+    'xml.searchLabel': 'Search the document',
+    'xml.scopeLabel': 'Search in',
+    'xml.scope.all': 'Everything',
+    'xml.scope.name': 'Names',
+    'xml.scope.value': 'Values',
+    'xml.scope.comment': 'Comments & PIs',
+    'xml.previousMatch': 'Previous match',
+    'xml.nextMatch': 'Next match',
+    'xml.matchPosition': '{current} / {total}',
+    'xml.noMatches': 'No matches',
+    'xml.matchesOnly': 'Matches only',
+    'xml.showAttributes': 'Attribute rows',
+    'xml.expandAll': 'Expand all',
+    'xml.collapseAll': 'Collapse all',
+    'xml.problems': '{count} notes',
+    'xml.problemsPartial': '{count} problems — the tree shows what could be recovered',
+    'xml.tree': 'Structure',
+    'xml.treeLabel': 'XML structure',
+    'xml.rowCount': '{count} rows',
+    'xml.inspector': 'Details',
+    'xml.source': 'Source',
+    'xml.lineCount': '{count} lines',
+    'xml.summary.elements': 'Elements',
+    'xml.summary.attributes': 'Attributes',
+    'xml.summary.depth': 'Max depth',
+    'xml.summary.namespaces': 'Namespaces',
+    'xml.summary.lines': 'Lines',
+    'xml.moreAttributes': '+{count} more',
+    'xml.childCount': '({count} children)',
+    'xml.kind.element': 'Element',
+    'xml.kind.attribute': 'Attribute',
+    'xml.kind.text': 'Text',
+    'xml.kind.cdata': 'CDATA',
+    'xml.kind.comment': 'Comment',
+    'xml.kind.pi': 'Processing instruction',
+    'xml.kind.doctype': 'Document type',
+    'xml.copyXpath': 'Copy XPath',
+    'xml.copyValue': 'Copy value',
+    'xml.copyXml': 'Copy XML',
+    'xml.showInSource': 'Show in source',
+    'xml.copied': 'Copied to the clipboard',
+    'xml.copyFailed': 'Copy failed',
+    'xml.document': 'Document',
+    'xml.namespaces': 'Namespace declarations',
+    'xml.noNamespaces': 'This document declares no namespaces.',
+    'xml.defaultNamespace': '(default)',
+    'xml.statistics': 'Statistics',
+    'xml.attributes': 'Attributes',
+    'xml.stat.text': 'Text nodes',
+    'xml.stat.cdata': 'CDATA sections',
+    'xml.stat.comments': 'Comments',
+    'xml.stat.pis': 'Processing instructions',
+    'xml.field.version': 'XML version',
+    'xml.field.encoding': 'Encoding',
+    'xml.field.encodingValue': '{used} (declared {declared})',
+    'xml.field.standalone': 'Standalone',
+    'xml.field.doctype': 'DOCTYPE',
+    'xml.field.publicId': 'Public ID',
+    'xml.field.systemId': 'System ID',
+    'xml.field.root': 'Root element',
+    'xml.field.namespace': 'Namespace',
+    'xml.field.prefix': 'Prefix',
+    'xml.field.uri': 'URI',
+    'xml.field.count': 'Declarations',
+    'xml.field.xpath': 'XPath',
+    'xml.field.position': 'Position',
+    'xml.field.positionValue': 'Line {line}, column {column}',
+    'xml.field.children': 'Children',
+    'xml.field.childrenValue': '{elements} elements · {attributes} attributes',
+    'xml.field.status': 'Status',
+    'xml.field.name': 'Name',
+    'xml.field.value': 'Value',
+    'xml.field.text': 'Text content',
+    'xml.unclosed': 'Not closed — its end tag is missing',
+    'xml.emptyText': 'No text content.',
+    'xml.truncated': 'Showing the first {count} characters.',
+    'diag.xml.noRoot': 'The file contains no XML element.',
+    'diag.xml.notText': 'The file is binary, not XML text.',
+    'diag.xml.unsupportedEncoding': 'The declared encoding {encoding} is not supported; the file was read as UTF-8.',
+    'diag.xml.mismatchedTag': 'End tag </{found}> does not match the open element <{expected}>.',
+    'diag.xml.unclosedElement': 'Element <{name}> is not closed.',
+    'diag.xml.endTagSyntax': 'End tag </{name}> may contain only its name.',
+    'diag.xml.missingAttributeSpace': 'Attribute {name} must be separated from the previous one by whitespace.',
+    'diag.xml.entityMarkup': 'Entity &{name}; contains markup, which is shown as text, not parsed into elements.',
+    'diag.xml.entityLtInAttribute': 'Entity &{name}; puts a \'<\' into an attribute value.',
+    'diag.xml.tagSyntax': 'Unexpected character {char} in tag <{name}>.',
+    'diag.xml.unexpectedEndTag': 'End tag </{name}> has no matching start tag.',
+    'diag.xml.duplicateAttribute': 'Attribute {name} appears more than once.',
+    'diag.xml.attributeSyntax': 'Attribute {name} needs a quoted value.',
+    'diag.xml.ltInAttribute': 'Attribute {name} contains a \'<\'.',
+    'diag.xml.undeclaredPrefix': 'Namespace prefix {prefix} is not declared.',
+    'diag.xml.multipleRoots': 'The document has more than one root element.',
+    'diag.xml.textOutsideRoot': 'Text appears outside the root element.',
+    'diag.xml.misplacedDeclaration': 'The XML declaration must be at the very start of the file.',
+    'diag.xml.misplacedDoctype': 'The DOCTYPE must come once, before the root element.',
+    'diag.xml.invalidMarkup': 'Unrecognized markup.',
+    'diag.xml.unescapedLt': 'A \'<\' that does not start markup must be written as &lt;.',
+    'diag.xml.bareAmpersand': 'An \'&\' that does not start a reference must be written as &amp;.',
+    'diag.xml.invalidCharRef': 'Invalid character reference {reference}.',
+    'diag.xml.undefinedEntity': 'Entity &{name}; is not declared and was left as written.',
+    'diag.xml.externalEntity': 'External entity &{name}; is not loaded and was left as written.',
+    'diag.xml.entityExpansionLimit': 'Entity expansion stopped at {max} characters; later references were left as written.',
+    'diag.xml.unterminatedTag': 'A tag is not closed with \'>\'.',
+    'diag.xml.unterminatedAttribute': 'An attribute value is missing its closing quote.',
+    'diag.xml.unterminatedComment': 'A comment is not closed with \'-->\'.',
+    'diag.xml.unterminatedCdata': 'A CDATA section is not closed with \']]>\'.',
+    'diag.xml.unterminatedPi': 'A processing instruction is not closed with \'?>\'.',
+    'diag.xml.unterminatedDoctype': 'The DOCTYPE declaration is not closed.',
+    'diag.xml.nodeLimit': 'Only the first {max} nodes were read.',
+    'diag.xml.depthLimit': 'Reading stopped at an element nested deeper than {max} levels; the rest of the document was not read.',
+    'diag.xml.moreDiagnostics': '{count} more problems are not listed.',
+    'epub.title': 'EPUB e-book',
+    'epub.eyebrow': 'EPUB {version}',
+    'epub.contents': 'Contents',
+    'epub.info': 'Book info',
+    'epub.previous': 'Previous',
+    'epub.next': 'Next',
+    'epub.chapter': 'Chapter',
+    'epub.position': '{position} / {total}',
+    'epub.textSmaller': 'Smaller text',
+    'epub.textLarger': 'Larger text',
+    'epub.textReset': 'Reset text size',
+    'epub.textScale': 'Text size {scale}% (reset)',
+    'epub.toolbar': 'Reading controls',
+    'epub.reader': 'Book text',
+    'epub.partial': 'Some parts of this book could not be read ({count} issues). See Book info.',
+    'epub.section': 'Section {number}',
+    'epub.untitled': 'Untitled',
+    'epub.cover': 'Cover',
+    'epub.moreDiagnostics': '{count} more issues not listed.',
+    'epub.diagnostics': 'Diagnostics',
+    'epub.blocked': '{count} external resources were blocked.',
+    'epub.meta.title': 'Title',
+    'epub.meta.otherTitles': 'Other titles',
+    'epub.meta.creators': 'Authors',
+    'epub.meta.contributors': 'Contributors',
+    'epub.meta.publisher': 'Publisher',
+    'epub.meta.date': 'Published',
+    'epub.meta.modified': 'Modified',
+    'epub.meta.language': 'Language',
+    'epub.meta.identifiers': 'Identifiers',
+    'epub.meta.series': 'Series',
+    'epub.meta.subjects': 'Subjects',
+    'epub.meta.description': 'Description',
+    'epub.meta.rights': 'Rights',
+    'epub.meta.source': 'Source',
+    'epub.meta.version': 'EPUB version',
+    'epub.meta.layout': 'Layout',
+    'epub.meta.direction': 'Page direction',
+    'epub.meta.chapters': 'Chapters',
+    'epub.meta.files': 'Files',
+    'epub.meta.package': 'Package document',
+    'epub.direction.rtl': 'Right to left',
+    'epub.direction.ltr': 'Left to right',
+    'epub.layout.reflowable': 'Reflowable',
+    'epub.layout.fixed': 'Fixed layout (shown reflowed)',
+    'epub.chapter.encrypted': 'This chapter is DRM-protected and cannot be shown.',
+    'epub.chapter.unsupported': 'This chapter is a {mediaType} document, which cannot be shown.',
+    'epub.chapter.missing': 'This chapter is listed in the book but missing from the file ({path}).',
+    'epub.chapter.renderFailed': 'This chapter could not be displayed ({path}).',
+    'epub.read.missing': 'This chapter is missing from the file ({path}).',
+    'epub.read.too-large': 'This chapter is larger than the configured read limit ({path}).',
+    'epub.read.unsupported-method': 'This chapter uses an unsupported ZIP compression method ({path}).',
+    'epub.read.encrypted': 'This chapter is encrypted ({path}).',
+    'epub.read.corrupted': 'This chapter is damaged and could not be decompressed ({path}).',
+    'epub.read.inflate-unavailable': 'This platform cannot decompress EPUB content.',
+    'epub.read.budget-exceeded': 'The decompression budget for this book has been used up.',
+    'diag.epub.not-zip': 'This file is not a readable EPUB (ZIP) container.',
+    'diag.epub.no-package': 'No package document (.opf) was found in this EPUB.',
+    'diag.epub.package-unreadable': 'The package document could not be read.',
+    'diag.epub.package-invalid': 'The package document is not a valid EPUB package.',
+    'diag.epub.no-spine': 'This EPUB has no readable reading order (spine).',
+    'diag.epub.drm': 'This EPUB is DRM-protected; its content cannot be shown.',
+    'diag.epub.inflate-unavailable': 'This platform cannot decompress EPUB content.',
+    'diag.epub.decompressed-limit': 'This EPUB declares more uncompressed data than can be safely opened.',
+    'diag.epub.zip-entry-limit': 'Only the first {limit} files in the container were read.',
+    'diag.epub.xml-truncated': 'An XML document was too large or deeply nested and was read partially.',
+    'diag.epub.mimetype-missing': 'The container has no mimetype file.',
+    'diag.epub.mimetype-mismatch': 'The container mimetype is "{value}" instead of application/epub+zip.',
+    'diag.epub.multiple-renditions': 'The book has {count} renditions; the first one is shown.',
+    'diag.epub.container-missing': 'META-INF/container.xml is missing or invalid; using {path}.',
+    'diag.epub.encoding': 'Unsupported text encoding "{encoding}" was read as UTF-8.',
+    'diag.epub.rights-file': 'The book contains a rights file (META-INF/rights.xml), which usually indicates DRM.',
+    'diag.epub.item-limit': 'Only the first {limit} items were read.',
+    'diag.epub.manifest-missing': '{count} files listed in the manifest are missing from the container.',
+    'diag.epub.spine-idref': 'The reading order references {count} unknown items (first: "{idref}").',
+    'diag.epub.spine-unsupported': 'A reading-order item of type {mediaType} cannot be displayed.',
+    'diag.epub.resource-missing': 'A content document listed in the reading order is missing.',
+    'diag.epub.encrypted': '{count} chapters are encrypted and cannot be shown.',
+    'diag.epub.toc-unreadable': 'The table of contents could not be read.',
+    'diag.epub.toc-missing': 'The book has no table of contents; chapters are listed in reading order.',
+    'diag.epub.time-limit': 'Parsing stopped at the time limit; the book may be incomplete.',
+    'diag.epub.fixed-layout': 'This is a fixed-layout book; pages are shown as reflowable text.',
+    'har.title': 'HAR network log',
+    'har.search': 'Search URL, status, headers…', 'har.searchBodies': 'Search bodies',
+    'har.version': 'HAR {version}', 'har.requestCount': '{count} requests', 'har.unknownCreator': 'Unknown writer',
+    'har.requests': 'Requests', 'har.pages': 'Pages', 'har.info': 'Archive info',
+    'har.rows': '{shown} / {total} rows shown', 'har.moreRows': '{count} more…',
+    'har.noMatches': 'No requests match the current filters.',
+    'har.noEntries': 'This archive contains no requests.',
+    'har.noPages': 'This archive records no page loads.',
+    'har.noTimings': 'This request has no timing breakdown.',
+    'har.selectRequest': 'Select a request to inspect it.',
+    'har.unreadable': 'This file could not be read as a HAR archive.',
+    'har.failed': 'failed', 'har.cached': 'cached', 'har.notApplicable': 'n/a',
+    'har.resourceSize': 'Resource size {size}', 'har.startedAt': 'started +{offset}',
+    'har.copyEntry': 'Copy request JSON', 'har.copyBody': 'Copy body',
+    'har.filter.method': 'Method', 'har.filter.allMethods': 'All methods',
+    'har.filter.status': 'Status', 'har.filter.allStatuses': 'All statuses',
+    'har.filter.type': 'Type', 'har.filter.allTypes': 'All types',
+    'har.filter.domain': 'Domain', 'har.filter.allDomains': 'All domains',
+    'har.statusClass.1xx': '1xx informational', 'har.statusClass.2xx': '2xx success',
+    'har.statusClass.3xx': '3xx redirect', 'har.statusClass.4xx': '4xx client error',
+    'har.statusClass.5xx': '5xx server error', 'har.statusClass.none': 'No response',
+    'har.type.document': 'Document', 'har.type.stylesheet': 'Stylesheet', 'har.type.script': 'Script',
+    'har.type.xhr': 'XHR / Fetch', 'har.type.image': 'Image', 'har.type.font': 'Font',
+    'har.type.media': 'Media', 'har.type.websocket': 'WebSocket', 'har.type.manifest': 'Manifest',
+    'har.type.text': 'Text', 'har.type.binary': 'Binary', 'har.type.other': 'Other',
+    'har.timing.blocked': 'Blocked', 'har.timing.dns': 'DNS', 'har.timing.connect': 'Connect',
+    'har.timing.ssl': 'TLS', 'har.timing.send': 'Send', 'har.timing.wait': 'Wait (TTFB)',
+    'har.timing.receive': 'Receive', 'har.timing.total': 'Total',
+    'har.column.index': '#', 'har.column.name': 'Name', 'har.column.method': 'Method',
+    'har.column.status': 'Status', 'har.column.type': 'Type', 'har.column.domain': 'Domain',
+    'har.column.size': 'Transferred', 'har.column.time': 'Time', 'har.column.waterfall': 'Waterfall',
+    'har.column.page': 'Page', 'har.column.title': 'Title', 'har.column.started': 'Started',
+    'har.column.contentLoad': 'DOMContentLoaded', 'har.column.load': 'Load', 'har.column.requests': 'Requests',
+    'har.detail.general': 'General', 'har.detail.headers': 'Headers', 'har.detail.payload': 'Payload',
+    'har.detail.response': 'Response', 'har.detail.cookies': 'Cookies', 'har.detail.timings': 'Timings',
+    'har.detail.rawTimings': 'Reported phases', 'har.detail.requestHeaders': 'Request headers',
+    'har.detail.responseHeaders': 'Response headers', 'har.detail.queryString': 'Query string',
+    'har.detail.requestBody': 'Request body', 'har.detail.responseBody': 'Response body',
+    'har.detail.requestCookies': 'Request cookies', 'har.detail.responseCookies': 'Response cookies',
+    'har.field.url': 'URL', 'har.field.method': 'Method', 'har.field.status': 'Status',
+    'har.field.protocol': 'Protocol', 'har.field.type': 'Type', 'har.field.started': 'Started',
+    'har.field.offset': 'Timeline offset', 'har.field.time': 'Duration', 'har.field.transferred': 'Transferred',
+    'har.field.resourceSize': 'Resource size', 'har.field.remoteAddress': 'Remote address',
+    'har.field.connection': 'Connection', 'har.field.redirect': 'Redirects to', 'har.field.page': 'Page',
+    'har.field.error': 'Error', 'har.field.comment': 'Comment', 'har.field.mimeType': 'MIME type',
+    'har.field.encoding': 'Encoding', 'har.field.compression': 'Saved by compression',
+    'har.body.none': 'None', 'har.body.empty': 'The archive stored no body text.',
+    'har.body.binary': 'Binary body ({size}) — stored as base64 and not decoded.',
+    'har.body.truncated': 'Body truncated at the preview limit.',
+    'har.body.incomplete': 'The stored body ends before its content does — part of it could not be decoded.',
+    'har.summary.requests': 'Requests', 'har.summary.elapsed': 'Elapsed', 'har.summary.transferred': 'Transferred',
+    'har.summary.resources': 'Resources', 'har.summary.domains': 'Domains', 'har.summary.errors': 'Errors',
+    'har.summary.pages': 'Pages',
+    'har.info.version': 'HAR version', 'har.info.creator': 'Writer', 'har.info.browser': 'Browser',
+    'har.info.entries': 'Requests', 'har.info.pages': 'Pages', 'har.info.domains': 'Domains',
+    'har.info.methods': 'Methods', 'har.info.firstRequest': 'First request', 'har.info.elapsed': 'Elapsed',
+    'har.info.totalTime': 'Sum of request times', 'har.info.transferred': 'Transferred',
+    'har.info.resources': 'Resource bytes', 'har.info.errors': 'Error responses',
+    'har.info.creatorComment': 'Writer comment',
     'coreml.title': 'Core ML model', 'coreml.search': 'Search operations, operands, and weights…',
     'coreml.copyJson': 'Copy JSON', 'coreml.graph': 'Graph', 'coreml.unnamed': 'unnamed',
     'coreml.operations': 'Operations', 'coreml.io': 'Inputs & Outputs', 'coreml.weights': 'Weights',
@@ -386,6 +734,7 @@ export const CATALOG_EN = {
     'pt2.warning.payloadsLimited': '{count} weight or constant entries were not read because the limit was reached.',
     'pt2.warning.danglingRefs': '{count} argument(s) reference a value that no node or input produces.',
     'pt2.warning.subgraphDepth': '{count} sub-graph(s) nested deeper than {limit} levels were not read.',
+    'pt2.warning.placeholdersLimited': 'Only the first {limit} graph inputs or outputs were read; {count} were dropped.',
     'pt2.warning.pickledPayloads': 'Program "{name}": {count} tensor(s) are stored as torch.save pickles and are listed without a value preview.',
     'pt2.warning.payloadMissing': 'Program "{name}": {count} weight file(s) referenced by the payload config are not in the archive.',
     'pt2.warning.payloadTruncated': 'Program "{name}": {count} weight file(s) are shorter than the tensor they hold.',
@@ -601,6 +950,18 @@ export const CATALOG_EN = {
     'diag.limit-exceeded.input': 'The file exceeds the maximum supported size ({maxBytes} bytes).',
     'diag.limit-exceeded.decompressed': 'The archive declares more uncompressed data than the supported limit ({maxBytes} bytes).',
     'diag.aborted': 'Parsing was cancelled.',
+    'diag.har.invalid': 'This file could not be parsed as JSON, so it cannot be read as a HAR archive.',
+    'diag.har.missing-log': 'This JSON file has no "log" object, so it is not a HAR archive.',
+    'diag.har.missing-entries': 'The HAR log has no "entries" array.',
+    'diag.har.no-entries': 'This archive records no requests.',
+    'diag.har.version': 'The archive declares HAR version {version}; it is read as 1.2.',
+    'diag.har.entry-limit': 'Only the first {shown} of {total} requests were read.',
+    'diag.har.invalid-entries': '{count} entries are not objects and were skipped.',
+    'diag.har.invalid-dates': '{count} request(s) have a start time that is not a valid ISO-8601 timestamp.',
+    'diag.har.body-truncated': '{count} body(ies) were truncated at {limit} characters.',
+    'diag.har.body-incomplete': '{count} response body(ies) could not be fully decoded.',
+    'diag.har.binary-bodies': '{count} response(s) hold binary bodies that are listed by size only.',
+    'diag.har.request-errors': '{count} request(s) failed before a response arrived.',
     'diag.csv.ragged-rows': '{count} row(s) had a different number of columns and were padded or truncated.',
     // JSON parser diagnostics (docs/parsers/json.md)
     'diag.limit-exceeded.nodes': 'Only the first {count} nodes were parsed (node limit reached).',