OPM Flow VS Code Extension

Language support for OPM Flow reservoir simulation deck files, with syntax highlighting and development features backed by the full OPM Flow reference manual.

Keyword data is built from OPM/opm-reference-manual at commit 52fbb530.

Features

Syntax Highlighting

Provides syntax highlighting for OPM Flow simulation deck files with support for:

• Section headers: RUNSPEC, GRID, EDIT, PROPS, REGIONS, SOLUTION, SUMMARY, SCHEDULE — scoped so most themes render them in a distinct color (yellow in Dark+)
• Keywords: ALL_CAPS identifiers (e.g., COMPDAT, WELSPECS, DATES)
• Comments: Lines starting with --
• Record terminators: / marking the end of a record
• Numbers: Integers and floating-point values
• Defaults / repeat markers: 1*, 3*, etc. (distinct from ordinary numbers)
• Strings: Text in single quotes
• Template variables: <NAME> placeholders used in macro/ERT workflows
• END keyword: Specially highlighted file terminator
• UDQ / ACTIONX constructs: UDQ control words (DEFINE, ASSIGN, UNITS, UPDATE), UDQ functions (SORTA, SUM, ABS, …), the comparison and logical operators used in expressions (>=, <=, ==, AND, OR), and the ACTIONX / ENDACTIO action-block delimiters

Keyword Autocompletion

Autocomplete support for OPM Flow keywords extracted from the reference manual. Each completion item shows the deck section(s) the keyword is valid in (RUNSPEC, GRID, SCHEDULE, …) and a one-line summary in the documentation pane. Keywords that are valid in every section (e.g. INCLUDE, ECHO) list them all. Completions are triggered when typing uppercase letters at the start of a line.

Accepting a completion inserts the keyword together with a boilerplate data record as a tab-navigable snippet: each parameter becomes a placeholder filled with its documented default, or a type-appropriate dummy value when there is no default (1 for integers, 0.0 for reals, 'STRING' for strings). The record is terminated to match the keyword's shape — a single / for fixed and array keywords, plus an extra standalone / line for record-list keywords like WELSPECS and COMPDAT. Activation keywords (e.g. OIL, UNIFOUT) insert the name alone. Press Tab to jump between placeholders and overwrite the dummy values. To insert just the keyword name with no record, set opm-flow.completion.keywordInsert to keyword (see Settings).

Parameter Value Completion

Inside a record, when the parameter at the current column has a known set of string options, the suggestion list shows them with a one-line description per option. For example, on a COMPDAT record the STATUS column offers OPEN, SHUT, and AUTO. Multi-record keywords resolve the parameter against the record the cursor is in, so each record gets the right option list.

To activate the list, position the cursor in the parameter column and either start typing an uppercase letter (e.g. O for OPEN) or press Ctrl+Space (Cmd+Space on macOS) to open the suggestions manually. Selecting an option inserts the value quoted, e.g. 'OPEN'.

When the current column is a well- or group-name parameter — recognised by the opm-common item name (the WELL/GROUP family and their spelling variants, so WCONPROD, WELOPEN, GCONPROD, WEFAC, … all qualify) — the list instead offers the names the deck already declares: well names from WELSPECS, and group names from WELSPECS (the parent-group column) and GRUPTREE. Names are gathered from the whole file, so opening a quote (') or pressing Ctrl+Space in a well column lists every well defined in the deck. The stringValueStyle setting controls quoting just as it does for option values.

UDQ and ACTIONX Support

The user-defined-quantity sub-language (UDQ blocks) and ACTIONX action blocks are recognised so the editor can assist with their distinct syntax:

• Completion — inside a UDQ block, the start of a statement offers the control words (ASSIGN, DEFINE, UNITS, UPDATE); inside a UDQ formula or an ACTIONX condition, the UDQ functions (SORTA, SUM, ABS, …) are offered and inserted with parentheses ready for the argument.
• Hover — hovering a UDQ control word shows what it does, and hovering a UDQ function shows its signature and description.
• Diagnostics — a UDQ statement that doesn't start with a control word, and an ACTIONX block left unclosed by ENDACTIO, are flagged (see Diagnostics).

Hover Tooltips

Hover over any keyword to see a quick tooltip with:

• The keyword name and all valid sections
• A summary from the reference manual
• A parameter table listing all record fields with units and defaults
• A usage example

Hovering over a value in a data record shows the description for that specific parameter column. For example, hovering over the group name in a WELSPECS record shows the GRPNAME parameter description, units, and default.

For a summary vector or array variant that derives from an opm-common keyword family — e.g. WOPR (from WELL_PROBE) or KRNUMX (from KRNUM) — the hover adds a Deck-name alias of <family> line, so you can see which keyword family the concrete name belongs to.

Keywords on the diagnostics exclusion list (see opm-flow.diagnostics.excludedKeywords under Settings) carry an extra notice in the hover indicating that arity, terminator, and section checks are skipped — useful when squiggles are deliberately suppressed for a keyword.

Diagnostics

Squiggles in the editor catch the most common deck-shape mistakes:

• Unrecognised keyword — a keyword-shaped line whose token isn't in the OPM Flow vocabulary (typo of an identifier, or a custom keyword the parser won't accept).
• Wrong section — a keyword used outside the section(s) it's valid in, e.g. WELSPECS placed in RUNSPEC instead of SCHEDULE. The hover lists the sections where it is valid.
• Over-arity record — a record with more values than the keyword's per-record item count from the OPM Flow parser. Trailing defaults (auto-defaulted by /) are not flagged. The squiggle starts at the first offending value, including N* repeat tokens.
• Value-type mismatch — a record value whose form is unambiguously wrong for the parameter's declared value_type, e.g. a quoted string in a numeric slot or a decimal in an integer slot. Deliberately conservative: bare identifiers (which may be UDA/UDQ references) and enum-option mismatches are never flagged, and defaults (*, N*) are always accepted.
• Missing per-record / — a record line carrying values but no closing /, on keywords known to take records.
• Missing closing / on record-list blocks — keywords like WELSPECS, COMPDAT, WCONHIST that expect the block to end with a standalone / line.
• Missing closing / on array blocks — cell-property arrays like PORO, PERMX, ACTNUM, OPERNUM, ZCORN. The closing / may sit on its own line or trail the last value line — both forms are accepted.
• Indented keyword — a recognised keyword that doesn't start in column 1. Per the OPM Flow reference manual, indented keywords are silently ignored by the simulator.
• Lowercase keyword — a line shaped like a keyword declaration whose upper-cased form is a known keyword. The reference manual states that lowercase keywords produce errors at simulation time.
• Missing required keyword — a keyword whose requires partner (from opm-common) is absent, e.g. a saturation-function table without its phase keyword. Suppressed when the deck pulls in other files via INCLUDE/IMPORT/GDFILE (the partner may live there) or has no section header (an include fragment, not a complete deck).
• Mutually exclusive keywords — two keywords that opm-common marks as prohibits partners both appearing in the same deck.
• UDQ statement without a control word — a statement inside a UDQ block that does not begin with ASSIGN, DEFINE, UNITS, or UPDATE. Continuation lines of a statement whose / is deferred are not flagged.
• Unclosed ACTIONX block — an ACTIONX action block with no matching ENDACTIO before the end of the deck.

Keywords whose record bodies don't fit the generic model can be silenced wholesale via the opm-flow.diagnostics.excludedKeywords setting — see Settings below.

Quick Fixes

Diagnostics that have an unambiguous correction offer a lightbulb Quick Fix (Ctrl+., or Cmd+. on macOS). Place the cursor on the squiggle, open the lightbulb, and apply:

• Convert to uppercase — a lowercase keyword like welspecs → WELSPECS.
• Move keyword to column 1 — strip the leading whitespace from an indented keyword.
• Add terminating / — append the missing per-record /.
• Add / to close the record list / value array — insert the missing standalone / line that closes a WELSPECS/COMPDAT block or a PORO/PERMX array.
• Replace with <nearest> — for an unrecognised keyword that is a close typo of a known one (e.g. EQLDIM → EQLDIMS), substitute the suggested keyword.

Docs Panel

Open the bottom panel (Ctrl+J) and select the OPM Keyword Reference tab. It updates automatically as you move the cursor — no keystrokes needed:

• Cursor on a keyword → full documentation: valid sections, summary, complete parameter table, example
• Cursor on a value column → same view with the matching parameter row highlighted
• Cursor on whitespace or a comment → panel retains the last shown keyword

The panel shows the keyword name, the section(s) it applies to, the summary, the parameter table, and the example. This is the main view for reading long keyword documentation, since it scrolls freely and stays visible while you edit. If you prefer it alongside the editor rather than beneath it, drag the tab into the primary or secondary sidebar — VS Code remembers the placement.

By default each parameter's Type, units, and Default are folded into a compact muted sub-line beneath its description (the embedded layout). This keeps the Description readable even for keywords with many parameters (e.g. WECON), where a wide column grid otherwise squeezes it into a sliver. Set opm-flow.docs.layout to columns to restore the original separate-column table; the per-column show/hide settings (see Settings) still apply in either layout.

Keyword Outline (Tree View)

The OPM Flow Outline panel in the Explorer sidebar shows the deck as a two-level tree — each section (RUNSPEC, GRID, EDIT, PROPS, REGIONS, SOLUTION, SUMMARY, SCHEDULE) with its keywords nested underneath. Click any keyword to jump straight to it in the editor; the tree selection also follows the cursor as you move through the deck, and the view tracks the active file. Sections are collapsible, so you can fold away the parts of a large deck you are not working on. Each node carries the keyword summary as a tooltip.

Collapse Sections and Keywords

Sections and individual keywords can be folded in the editor gutter. A section runs from a section keyword (RUNSPEC, GRID, EDIT, PROPS, REGIONS, SOLUTION, SUMMARY, SCHEDULE) until the next section keyword, END, or end of file. Individual keyword folds nest inside their section so you can collapse whole sections at once or drill in one keyword at a time.

Align Record Columns

Tidy up record blocks so every column lines up. Three levels are available from the Command Palette or the editor right-click menu, depending on how much you want to touch:

• OPM Flow: Align Record Columns in Record — aligns only the record group under the cursor.
• OPM Flow: Align Record Columns in File — aligns the whole current file, or just the selected lines when there is a selection.
• OPM Flow: Align Record Columns in Deck — follows the INCLUDE chain and aligns every reachable file (see below).

Groups of consecutive record lines (same token count) are reformatted in place: strings left-aligned, integer columns right-aligned, and float columns aligned on the decimal point (integers and N* repeat markers in a float column line up at the decimal point position). Keyword headers, comment lines, the closing /, and trailing -- comments are left untouched.

Each row is given a fixed leading indent: opm-flow.formatting.recordIndent (default 2) for a plain table, or opm-flow.formatting.headingIndent (default 3) when the table has a column heading — see Settings. A single-row table is aligned as well: it has no other rows to line up against, but it still gets the configured indent and single-space column separation (and stays in sync with a heading above it).

-- comment lines interspersed within a record group no longer break the group — every data line above and below the comment is aligned against a single shared set of column widths.

Comments are ignored when aligning, with one exception: a column heading directly above the group — a -- comment with exactly one word per column, as produced by Add Column Headers — is honoured. The records are aligned to it and the heading is kept in sync, so a table stays lined up under its heading across repeated alignments. Every other comment line is left exactly as it is: a descriptive comment above a table (any other word count, or one separated by a blank line) is never mistaken for a heading, and comments interspersed within the group are untouched.

Before:

MULTIPLY
 'PERMZ' 0.2 1 24 1 62 1 1 /
 'PERMZ' 0.04 1 24 1 62 2 2 /
 'PERMZ' 0.016 1 24 1 62 18 18 /
 'PERMZ' 1 1 24 1 62 22 22 /
/

After:

MULTIPLY
 'PERMZ' 0.2   1 24 1 62  1  1 /
 'PERMZ' 0.04  1 24 1 62  2  2 /
 'PERMZ' 0.016 1 24 1 62 18 18 /
 'PERMZ'     1 1 24 1 62 22 22 /
/

Align across the whole deck

OPM Flow: Align Record Columns in Deck starts from the active document, follows every INCLUDE (resolving PATHS aliases) recursively, and aligns the record tables in all reachable files in one operation. It reports how many lines and files were changed when it finishes.

When sweeping the whole deck, per-cell grid/region/solution arrays (PORO, PERMX, COORD, SATNUM, …) and large tables (VFPPROD/VFPINJ) are skipped by default so an INCLUDEd grid file is not silently rewritten and its deliberate fixed-width layout is preserved. The in Record and in File commands do not apply these defaults — when you align text you explicitly targeted, everything is aligned.

Add further keywords to skip via opm-flow.formatting.alignColumnsExcludedKeywords (honoured by all three commands, and added to the deck defaults).

Add Column Headers

Invoke OPM Flow: Add Column Headers from the Command Palette or the right-click menu to insert a -- comment above the record group with parameter names taken from the keyword documentation, then align the records to those positions.

Existing comments around the table are ignored, so a descriptive comment above the data is never mistaken for a heading. If the line directly above the group is a heading this command previously generated (its words are exactly the column names), it is updated in place; otherwise a new heading line is inserted. Running the command multiple times is idempotent.

Example — cursor anywhere inside the VFPIDIMS record:

VFPIDIMS
-- MXMFLO MXMTHP MXVFPTAB
      30     20       20 /

Toggle Line Comment

Select one or more lines and invoke OPM Flow: Toggle Line Comment (bound to Ctrl+/, or Cmd+/ on macOS — VS Code maps this to the same physical key as the built-in comment toggle, so it follows your keyboard layout) — also available from the right-click menu — to add or remove a -- comment marker at the very start of each selected line. If every non-blank line in the selection is already commented the marker is removed; otherwise -- is inserted at column 0 of each line. Blank lines are left untouched.

-- WCONPROD
--   'PROD' 'OPEN' /

File Navigation (INCLUDE / IMPORT / RESTART / GDFILE)

Quoted file paths on INCLUDE, IMPORT, RESTART, and GDFILE statements become clickable document links. Hold Ctrl (or Cmd on macOS) and click the path — or right-click and choose Go to Definition — to open the referenced file. Paths are resolved relative to the including file's directory, and any PATHS aliases ($NAME lookups) defined in the deck are expanded before the file is resolved.

INCLUDE
  'grid/PERM.inc' /

Open PRT File

With a .DATA deck open, OPM Flow: Open PRT File opens the matching <CASE>.PRT print/output file from the same folder — the file OPM Flow writes after a run. It is available from the editor right-click menu, the Command Palette, or the Ctrl+Alt+P (Cmd+Alt+P on macOS) shortcut. If the deck has not been run yet and no .PRT exists beside it, a notice is shown.

Verify and Run the Simulation (optional)

When a local OPM Flow binary is configured, two commands let you act on the open deck without leaving the editor — both available from the editor right-click menu and the Command Palette:

• OPM Flow: Verify Deck (dry run) runs flow in dry-run mode. The deck is fully parsed and the model is initialized (grid, properties, wells, schedule), but the time steps are not simulated — a fast check that the deck and everything it pulls in via INCLUDE / PATHS loads cleanly.
• OPM Flow: Run Simulation runs the full simulation.

Both run in an integrated terminal, started in the deck's own directory so its relative paths resolve. The output appears live, and afterwards Open PRT File shows the print file.

This is opt-in: nothing runs until you set opm-flow.simulator.executablePath (see Simulator under Settings). On Windows, OPM Flow is typically installed inside WSL — enable opm-flow.simulator.useWsl and point the executable path at the Linux binary (e.g. /usr/bin/flow); the deck's Windows path is translated to its /mnt/<drive> mount automatically.

Generate Keyword Reference

OPM Flow: Generate Keyword Reference (Command Palette Ctrl+Shift+P) opens a Markdown document listing all keywords grouped by section — useful for uploading as context to an AI chat session.

Settings

Configure via File → Preferences → Settings and search for opm-flow, or edit settings.json directly. All settings are scoped per-resource so you can override them per-workspace or per-folder.

Diagnostics

Setting	Default	Description
opm-flow.diagnostics.excludedKeywords	["RPTSCHED"]	Keywords to skip in every diagnostic check. Names are upper-cased on read; matching is case-insensitive. Add keywords whose record bodies don't fit the generic model and produce noisy false positives.

File associations

Setting	Default	Description
opm-flow.additionalFileExtensions	[]	Extra file extensions (with or without a leading .) to open as OPM Flow on top of the built-in list. Useful for project-specific include-file conventions. Matched case-insensitively. Example: [".myinc", "wellconv"]. For one-off cases the VS Code-native files.associations setting still works too.

Formatting

Setting	Default	Description
opm-flow.formatting.alignColumnsExcludedKeywords	[]	Extra keywords to skip when aligning record columns. Names are upper-cased on read; matching is case-insensitive. Honoured by all three Align Record Columns commands, and added to the built-in deck-sweep defaults (per-cell grid/region arrays like PORO/PERMX/COORD and large VFPPROD/VFPINJ tables, which Align … in Deck already skips).
opm-flow.formatting.recordIndent	2	Number of spaces placed before each record row when aligning a table that has no column heading. Applied by every Align Record Columns command (and to UDQ statement blocks).
opm-flow.formatting.headingIndent	3	Number of spaces placed before each record row when the table has a column heading (a -- comment with one label per column). Applied by Add Column Headers and when alignment keeps a table lined up under its heading.

Simulator

Optional integration for the Verify and Run commands. Leave executablePath unset to keep the feature dormant.

Setting	Default	Description
opm-flow.simulator.executablePath	"flow"	Path to the OPM Flow executable. With useWsl enabled this is a path inside WSL (e.g. /usr/bin/flow); otherwise a native path or a name on PATH.
opm-flow.simulator.useWsl	false	Run flow through WSL (wsl.exe). Required on Windows. The deck's Windows path is translated to its /mnt/<drive> mount automatically.
opm-flow.simulator.wslDistribution	""	WSL distribution to use (e.g. ubuntu-26.04); empty uses the default. List installed distributions with wsl -l -v. Only used when useWsl is enabled.
opm-flow.simulator.verifyArgs	["--enable-dry-run=true"]	Arguments for Verify Deck. The default parses and initializes the model without simulating time steps. Add --parsing-strictness=high to also fail on unsupported keywords.
opm-flow.simulator.runArgs	[]	Extra arguments for Run Simulation (e.g. --output-dir=..., --threads-per-process=4). The deck file is supplied automatically.

Completion

Setting	Default	Description
opm-flow.completion.keywordInsert	"template"	What accepting a keyword completion inserts. "template" adds the keyword plus a boilerplate data record (typed placeholders / documented defaults) as a tab-navigable snippet; "keyword" inserts just the keyword name.
opm-flow.completion.stringValueStyle	"quoted"	How STRING-typed parameter values appear in the suggestion list. "quoted" shows only 'OPEN'; "unquoted" shows only OPEN; "both" shows each option twice (e.g. OPEN and 'OPEN'). Inside an existing quoted token only the quoted form is offered regardless of this setting.

Docs layout

Setting	Default	Description
opm-flow.docs.layout	"embedded"	How the docs panel and hover tooltips lay out parameter metadata. "embedded" folds Type, units, and Default into a compact sub-line beneath each description (more room for the description); "columns" renders them as separate table columns (the original layout). The show/hide settings below apply in either layout.

Docs panel and hover columns

These toggles control which Type, unit, and Default values appear in the keyword docs panel and hover tooltips — as columns in the columns layout, or as bits of the metadata sub-line in the embedded layout. Disabling unused values gives the description more horizontal room when the view is docked somewhere narrow.

Setting	Default	Description
opm-flow.columns.showType	true	Show the parameter Type column (value type and dimension).
opm-flow.columns.showDefault	true	Show the Default column.
opm-flow.units.showField	true	Show the Field unit column.
opm-flow.units.showMetric	true	Show the Metric unit column.
opm-flow.units.showLab	true	Show the Laboratory unit column.

Supported File Extensions

The extension activates for the following extensions (case-sensitive on some platforms — both common casings are registered where relevant):

Core deck files: .data, .DATA, .dat, .inc, .INC, .incl, .include, .sch, .SCH, .sched, .schedule, .summary, .smry, .grdecl, .GRDECL, .grid, .gridopts, .vfp, .VFP, .vfpprod, .prop, .prpecl, .Ecl, .ecl.

Section data files (Eclipse/OPM include conventions): .aqucon, .aqunum, .dimens, .eqldims, .eqlnum, .equil, .fault, .faults, .fipnum, .fipzon, .multnum, .multregp, .multregt, .nnc, .ntg, .opernum, .perm, .permx, .poro, .pvt, .pvtnum, .regdims, .rocknum, .rxvd, .satnum, .sattab, .swatinit, .tabdims, .thpres, .trans.

For project-specific extensions not covered by this list, set opm-flow.additionalFileExtensions (see Settings).

Language ID

The language is registered as opm-flow.

Requirements

• VS Code 1.74.0 or later
• Python 3.10+ with lxml (only required when regenerating the keyword index)

Release Notes

0.9.0

• Keyword reference moved to the bottom panel — the OPM Keyword Reference webview now opens as a tab in the bottom panel (Ctrl+J) instead of its own activity bar container. Drag the tab into either sidebar if you prefer it beside the editor; VS Code remembers the placement.
• Align Record Columns at three scopes — the alignment command is now split into Align Record Columns in Record (the group under the cursor), … in File (the whole document, or the selection), and … in Deck (follows the INCLUDE chain, resolving PATHS aliases, and aligns every reachable file, reporting the line/file counts). The deck sweep skips per-cell grid/region arrays (PORO, PERMX, COORD, …) and large VFPPROD/VFPINJ tables by default so INCLUDEd fixed-width files are not rewritten; the new opm-flow.formatting.alignColumnsExcludedKeywords setting adds further keywords to skip (honoured by all three commands).
• Alignment indent and heading handling — the indent applied to record rows and to heading comments is now configurable via opm-flow.formatting.recordIndent and opm-flow.formatting.headingIndent. A group preceded by a -- column-heading comment aligns to that heading's columns rather than to the data alone, and single-row tables are aligned instead of left as written.
• Keyword completion no longer returns an empty list — ten curated supplemental keywords (STORE, AIM, CVTYPE, AMF, …) were merged into the runtime index without a sections array, and call sites that read it unconditionally threw part-way through building the suggestion list. Hitting any of them aborted the whole list, so keyword completion silently offered nothing.
• Well / group name completion — in a well- or group-name column (the opm-common WELL/GROUP item family — WCONPROD, WELOPEN, GCONPROD, WEFAC, …) the suggestion list offers the names the deck already declares: well names from WELSPECS, and group names from WELSPECS and GRUPTREE, gathered across the whole file. Each suggestion follows the quote style the name was declared with — a name written 'P1' is offered quoted, a bare name unquoted — and the list is sorted naturally, so W2 precedes W10.
• Verify and run the simulation (optional) — new OPM Flow: Verify Deck (dry run) and OPM Flow: Run Simulation commands launch a locally installed flow binary on the open deck in an integrated terminal. Verify uses flow's dry-run mode to confirm the deck (and its INCLUDE / PATHS files) loads without solving. Configure via the new opm-flow.simulator.* settings; Windows users can target a WSL distribution, with deck paths translated to /mnt/<drive> automatically.
• UDQ and ACTIONX support — the UDQ expression sub-language and ACTIONX action blocks are now recognised. Syntax highlighting scopes UDQ control words, UDQ functions, and expression operators, plus the ACTIONX / ENDACTIO block delimiters. Hover and completion cover UDQ control words and functions, and two new diagnostics flag a UDQ statement that doesn't start with a control word and an ACTIONX block left unclosed by ENDACTIO. Column alignment also gives UDQ expression blocks a dedicated three-column layout (control word, variable, expression) that treats a / used for division as part of the expression rather than the record terminator.
• Boilerplate keyword completion — accepting a keyword completion now inserts a sample data record as a tab-navigable snippet (documented defaults or type-appropriate dummy values, terminated to match the keyword's shape) instead of just the keyword name. The new opm-flow.completion.keywordInsert setting switches back to name-only insertion.
• Diagnostic quick fixes — lightbulb fixes for common deck mistakes: uppercase a lowercase keyword, move an indented keyword to column 1, add a missing record / list / array terminator /, and replace an unrecognised keyword with its nearest known match.
• Value-type diagnostic — record values are now checked against each parameter's declared opm-common type; an unambiguous mismatch (a quoted string in a numeric slot, a decimal in an integer slot) is flagged, while bare identifiers and enum-option mismatches are left alone to avoid false positives.
• Cross-keyword requires / prohibits diagnostics — flags a keyword whose required partner is missing from the deck, and a pair of mutually-exclusive keywords that both appear. The requires check is suppressed for include-based decks and bare include fragments.
• Fewer false positives on valid decks — broadened keyword recognition and relaxed terminator rules, validated against the OPM/opm-tests corpus:
  • ROCK is no longer reported as missing a list terminator. opm-common gives it the sentinel size SPECIAL_CASE_ROCK, which was classified as an unbounded list and so demanded a standalone / that no deck writes — the warning fired on roughly a quarter of the corpus.
  • TEMPVD is accepted in SOLUTION, and without a standalone /. opm-common defines it in two dialects, and only the narrower Eclipse100 definition was kept, dropping both the SOLUTION section and the EQLDIMS-dependent record count.
  • More SUMMARY vectors are recognised rather than flagged as unknown keywords: PROBE deck_names, directional KRNUM/IMBNUM region variants, L-modifier vectors (WOPRL, LWWIR), FIP region-set vectors (ROIP_ABC), UDQ names (WUOPRL, FU_VAR1), and resolved deck_name aliases.
  • Variadic-record list keywords (VFPPROD, VFPINJ, RSVD) no longer require a separate closing /.
  • The missing-record-terminator check is deferred until a record is actually left unterminated, so a / placed on a later line is accepted.
  • The wrong-section check is suppressed after an INCLUDE, whose file may supply the section header.
  • An indented unknown token under an active keyword is treated as record body rather than flagged as an unknown keyword.
• Deck-name alias resolution — concrete summary vectors and directional array variants now record which opm-common keyword family they derive from (e.g. WOPR → WELL_PROBE, KRNUMX → KRNUM) and surface it in hover. The 31 family/container schema names themselves (WELL_PROBE, AQUIFER_PROBE_ANALYTIC, MULT_XYZ, …) are no longer offered as completions or accepted as valid keywords, since they are never typed in a deck.

0.8.0

• Keyword outline tree view (Issue #41) — a new OPM Flow Outline panel in the Explorer shows sections and their keywords as a navigable, collapsible tree. Click a keyword to jump to it in the editor; the selection follows the cursor and the view tracks the active file.
• Compact parameter docs layout (Issue #26) — parameter Type, units, and Default are now folded into a muted sub-line beneath each description (the new default embedded layout), giving long descriptions far more room in the docs sidebar and hover tooltips. Set opm-flow.docs.layout to columns to keep the previous separate-column table.
• Open PRT file (Issue #44) — open the matching <CASE>.PRT print file next to a .DATA deck from the editor right-click menu, the Command Palette, or the new Ctrl+Alt+P (Cmd+Alt+P on macOS) shortcut.
• Section headers with trailing separators (Issue #31) — a decorated section line such as GRID ============== is now recognised as a section header, so keywords after it are no longer wrongly flagged "not valid in RUNSPEC" and section folds still open correctly.
• Dependency and CI maintenance — routine dev-dependency and GitHub Actions updates.

0.7.1

• Toggle line comment (Issue #27) — comment/uncomment the current line or selection with the standard editor shortcut, using OPM Flow's -- comment syntax.
• Refreshed keyword data — the bundled OPM Flow reference manual and opm-common keyword definitions were updated to their latest upstream revisions, broadening keyword coverage, section validity, and per-parameter value-type/dimension information surfaced in hover and the docs sidebar.
• Security hardening — CI workflow permissions and action pinning, plus webview/hover and policy hardening.

0.6.4

• Issue #13 — Unquoted string values like YES under SCALECRS or THPRES under EQLOPTS are no longer mis-classified as keywords. The grammar is anchored to column 1; the analyzer treats indented keyword- shaped tokens inside an open record block as record values; cursor-driven docs/hover require the word to start at column 0 before treating it as a keyword declaration.
• Bare uppercase tokens inside records are now coloured as strings. Well/group/property names like OP01, FIELD, UPPER, SGL/SOWCR in EQUALS/GRUPTREE/WCONHIST records no longer pick up the keyword colour. The keywords rule was tightened to match only when the line is the keyword alone (optionally followed by a -- comment or a /).
• Multi-line records no longer false-flag as missing /:
  • MESSAGES 13-INT records split as print-limits then stop-limits.
  • VFPPROD / VFPINJ axis (LIQ/THP/WFR/GFR/ALQ) and BHP tables; the block no longer demands a closing standalone / either.
  • WOPR, WGPR and 976 other non-F SUMMARY mnemonics with names spread across multiple lines.
• Bare-stacked SUMMARY mnemonics like GMWPR \n GMWIN \n / are accepted (new optional_body shape tagged on 981 non-F mnemonics so empty bodies don't demand a closing /; once values are listed the diagnostic still fires).
• Keyword recognition broadened:
  • SUMMARY-section mnemonics (FOPR, WOPR, GGOR, GPR, …).
  • Templated tracer mnemonics (FTPRSEA, WTITHTO, …) and TVDP mnemonics (TVDPFWT1, TVDPSIGS, …) resolved via base-name + suffix lookup.
  • OPM Flow Performance mnemonics (TCPUDAY, ELAPSED, …) and Control Mode Reporting mnemonics (FMCTP, WSTAT, …).
  • User-defined FIP region keywords (FIPZON, FIPGL, FIPNL, FIPUNIT, FIPHC, …).
  • UDQ SUMMARY mnemonics (WUWI1, FUOIL, GUTOT, …) via 2-char scope- prefix templates (FU, WU, GU, CU, RU, SU).
• File links extended from INCLUDE only to also cover IMPORT, RESTART, and GDFILE. PATHS aliases ($NAME lookups) are expanded before resolution.
• Variadic-record keyword continuation lines (RSVD, RVVD, PVDO, PVTO) no longer flagged as missing per-record /.
• TITLE no longer requires a trailing /.
• Multi-record keywords with integer size (TUNING, TUNINGL, TUNINGS, PLYSHLOG, PRORDER, PYACTION) are classified as fixed, not list, so no spurious closing-/ warning.
• 17 additional file extensions registered: .dat, .eqldims, .faults, .fipzon, .grid, .gridopts, .incl, .permx, .prpecl, .pvtnum, .regdims, .rxvd, .sched, .smry, .swatinit, .trans, .vfpprod.
• New opm-flow.additionalFileExtensions setting for project-specific include-file extensions not covered by the built-in list. Resource-scoped so a workspace can pin its own list.

0.6.3

• New diagnostic: keyword not starting in column 1. Per the OPM Flow reference manual, indented keywords are silently ignored by the simulator.
• New diagnostic: keyword in non-uppercase form. Lines shaped like a keyword declaration whose upper-cased form is a recognised keyword are flagged because OPM Flow only accepts uppercase keywords.
• Hover on a keyword that is on the diagnostics exclusion list (opm-flow.diagnostics.excludedKeywords) now shows a notice explaining that arity, terminator, and section checks are skipped for that keyword.
• New opm-flow.completion.stringValueStyle setting controls how STRING-typed parameter values appear in completions: quoted (default), unquoted, or both.
• Free-form text after a record-terminating / is now treated as a trailing comment (the -- prefix is no longer required).
• Bare uppercase identifiers inside an open record block are recognised as unquoted string values rather than flagged as unknown keywords (e.g. an INCLUDE followed by an unquoted path on the next line).

0.6.2

• Multi-record keyword support: hover, docs panel and arity diagnostics resolve the parameter against the record the cursor is in for keywords like WELSEGS, VFPPROD, COMPSEGS, ACTIONX, TUNING.
• Diagnostics extended to flag unknown keywords, missing per-record / terminators, and missing closing / on record-list and array blocks.
• New opm-flow.diagnostics.excludedKeywords setting to silence diagnostics for keywords whose record bodies don't fit the generic model.
• New column-toggle settings (opm-flow.columns.*, opm-flow.units.*) to hide unused columns in the docs sidebar and hover tooltips.
• Keyword data merged from the opm-common submodule, with corrected arity classification.
• Recover keywords with duplicate xml:id from the manual and parse dual-name PVT parameter rows.