Diagnostic code catalog

Every diagnostic rule-line reports carries a stable code. The code is the contract: integrations branch on it and tests assert on it, never on the message text. Message text is display copy and may change — including being translated — while the code stays fixed.

Some codes are reserved: registered now so the contract is stable, but not yet emitted by any producer. Each is marked "Reserved — not yet emitted", so a code with no current thrower is not mistaken for a missing feature.

A code marked "Retired — no longer emitted" was reported by an earlier version and no longer is. It stays registered so that an older record which carries it can still be read and explained.

This page is generated from DIAGNOSTIC_CODES, imported through the engine's published @one-unity/rule-line-engine/analysis subpath — adding a code without regenerating this page fails docs:check.

Code	Description	Status
FIXTURE_CONTRACT_DRIFT	The fixture input or expected output differs from the ruleset’s declared schema. Review the calling system’s contract; this warning does not prevent saving the fixture.	Active
FIXTURE_RECORDING_MISSING	This fixture has no recorded model response for the current document, prompt, schema and model. Re-record the response before using this fixture.	Active
FIXTURE_SNAPSHOT_TOO_LARGE	This fixture’s inline dataset snapshot exceeds the permitted byte size. Reduce it or pin a stored dataset version.	Active
FIXTURE_SNAPSHOT_SET_TOO_LARGE	The inline dataset snapshots in this fixture set exceed the combined permitted byte size. Reduce them or pin stored dataset versions.	Active
FIXTURE_IGNORE_PATH_INVALID	An ignored fixture path is invalid. Use member names, array indexes or [*] for one array level, starting with a member or an array index.	Active
CHECK_RULE_CODE_DUPLICATE	Two or more check steps use the same rule code. Give each business rule its own code so a collected violation identifies the rule that raised it.	Active
LOCALIZED_STRING_LANGUAGE_MISSING	A check message is missing non-empty German or English text. Template publication requires both languages; tenant messages may use a single language.	Active
DSL_INVALID_DOCUMENT	The submitted ruleset document is not a valid object and cannot be read at all.	Active
DSL_UNKNOWN_STEP_TYPE	A step in this ruleset uses a type that this version of rule-line does not recognize.	Active
DSL_MISSING_REQUIRED_FIELD	A field that this ruleset requires is missing from the document.	Active
DSL_INVALID_TYPE	A field in this ruleset has a value of the wrong kind, such as text where a number was expected.	Active
DSL_UNSUPPORTED_SCHEMA_VERSION	This ruleset declares a schema version that this version of rule-line does not support.	Active
DSL_DUPLICATE_STEP_SLUG	Two steps in this ruleset share the same identifying name, which must be unique.	Active
DSL_INVALID_STEP_SLUG	A step's identifying name does not match the naming pattern this ruleset requires.	Active
DSL_EMPTY_STEP_LIST	This ruleset does not declare any steps, but at least one step is required to run it.	Active
ENV_LIMIT_EXCEEDED	A run of this ruleset used more of a declared resource than its envelope allows.	Reserved — not yet emitted
ENV_DECLARED_EXCEEDS_PLATFORM_MAXIMUM	This ruleset declares a resource limit higher than the platform allows for any tenant.	Active
ENV_FIELD_INHERITS_MAXIMUM	This ruleset does not declare its own limit for a resource, so the platform's maximum applies instead.	Active
AI_SPEND_CAP_EXCEEDED	This run was stopped because it would have spent more on AI calls than the tenant's cap allows.	Reserved — not yet emitted
EXPR_UNKNOWN_FUNCTION	A formula in this ruleset calls a function name that does not exist in rule-line's function library.	Active
EXPR_SYNTAX_ERROR	This formula could not be read as valid expression syntax.	Active
EXPR_WRONG_ARITY	A function in this formula was called with the wrong number of arguments.	Active
EXPR_TYPE_MISMATCH	A value in this formula is not the kind of value an operator or function expects at that position.	Active
EXPR_DIVISION_BY_ZERO	This formula divides by zero, which has no defined result.	Active
EXPR_UNDEFINED_ARITHMETIC	This formula performs arithmetic on a value that is missing or null, which has no defined result.	Active
EXPR_INVALID_ROUNDING_MODE	This formula names a rounding mode that rule-line does not recognize.	Active
EXPR_CONTEXT_NOT_REPRESENTABLE	A value supplied as data for this run cannot be represented by the engine, so the value was refused.	Active
ENGINE_INTERNAL_ERROR	An unexpected fault occurred inside the engine itself. This is not feedback about your ruleset or formula.	Active
PATH_SYNTAX_ERROR	A data-access path in this formula could not be read as valid syntax.	Active
PATH_ABSENT	A data-access path in this formula did not resolve to a value for the given data.	Active
EXPR_NULL_IN_ARITHMETIC	This formula performs arithmetic on a value that is explicitly null, which has no defined result.	Active
EXPR_IMPLICIT_CONVERSION_REJECTED	This formula would need to convert text into a number implicitly, which rule-line never does automatically. Use an explicit conversion function instead.	Reserved — not yet emitted
EXPR_PLUS_ON_NON_NUMBER	The '+' operator only adds numbers in rule-line. To join text together, use CONCAT instead.	Active
EXPR_INCOMPARABLE_ORDERING	This formula orders two values of kinds that cannot be meaningfully compared to each other.	Active
EXPR_CONDITION_NOT_BOOLEAN	A condition in this ruleset must evaluate to true or false, but this one evaluates to a different kind of value.	Active
EXPR_NUMBER_PARSE_FAILED	This text could not be read as a number in the format that was specified.	Active
EXPR_COMPLEXITY_LIMIT_EXCEEDED	This formula is more deeply nested, or longer, than rule-line allows.	Active
EXPR_EMPTY_AGGREGATION	This formula aggregates a collection with no applicable values, which has no defined result.	Active
EXPR_GROUP_KEY_NOT_STRING	Grouping in this formula requires a text key, and this key is not text.	Active
EXPR_NO_MATCH	This formula expected to find exactly one matching element, but none of the elements matched.	Active
EXPR_MULTIPLE_MATCHES	This formula expected to find exactly one matching element, but more than one element matched.	Active
EXPR_COALESCE_EXHAUSTED	Every value given to this fallback function was absent or null, so it has no result to fall back to.	Active
EXPR_JOIN_NON_STRING	This formula joins values that are not all text. Convert each value to text explicitly first.	Active
EXPR_STRING_INDEX_OUT_OF_RANGE	This formula accesses a position in a piece of text that does not exist.	Active
EXPR_INVALID_DATE_PATTERN	This date pattern uses a token that rule-line does not recognize.	Active
EXPR_DATE_PARSE_FAILED	This text could not be read as a date using the pattern that was given.	Active
EXPR_UNKNOWN_TIMEZONE	This ruleset's declared timezone is not one rule-line recognizes.	Active
EXPR_MISSING_PRESENT_INSTANT	A date/time function needing the current moment was evaluated without one available.	Active
EXPR_LOCAL_TIME_NONEXISTENT	This local date/time does not exist in the ruleset's declared timezone, because it was skipped by a daylight-saving change; it was moved forward to the next time that does exist, and the run continued.	Active
PATH_INDEX_OUT_OF_RANGE	This data-access path reaches for a position in a collection that is beyond its bounds.	Active
PATH_INDEX_NOT_INTEGER	This data-access index or bound is not a whole number, so it cannot name a position.	Active
PATH_TARGET_NOT_ARRAY	This data-access path applies array-only syntax to a value that is not an array.	Active
RE_INVALID_PATTERN	This regular expression pattern could not be read as valid syntax.	Active
RE_DISALLOWED_CONSTRUCT	This regular expression uses a construct outside the pattern language rule-line supports: a lookahead or lookbehind, a back-reference, a named group, a character outside the Basic Multilingual Plane, a repeat count above 1000, an escape other than a character class shorthand or an escaped punctuation mark (such as \n, \t or \x41), or an unescaped {.	Active
RE_DISALLOWED_FLAG	This regular expression uses a flag that rule-line does not permit.	Active
RE_INPUT_TOO_LONG	The text given to this regular expression is longer than rule-line allows a pattern to be applied to.	Active
RE_ANCHOR_IN_WHOLE_MATCH	This regular expression already anchors itself, and rule-line cannot safely add its own anchoring around it.	Active
RE_ADJACENT_OVERLAPPING_QUANTIFIERS	Retired: this code is no longer reported. Regular expressions now run on a linear-time matcher, so repeated parts that can match the same characters no longer take long on long text. It stays registered so that older records which carry it can still be read.	Retired — no longer emitted
RE_PATTERN_TOO_LONG	This regular expression is too long as written, or too large once its repeat counts are expanded. While the ruleset is a draft this is a warning; the ruleset cannot be activated, and the pattern is never run, until the pattern is shortened.	Active
PATH_STEP_REFERENCE_UNRESOLVED	This formula references another step whose output could not be resolved in this ruleset.	Active
EXPR_FALLBACK_TYPE_MISMATCH	A step's declared fallback value does not match that step's declared output type.	Active
TYPE_STATIC_MISMATCH	A formula uses a value of the wrong kind, or the declared output schema does not accept what the output step produces, and this is knowable before the ruleset ever runs.	Active
TYPE_UNKNOWN_FIELD	A formula reads a field that the value it reads from does not have, whether that is the ruleset's input, a document, a step's output or a dataset row.	Active
TYPE_PATH_NOT_NAVIGABLE	A formula navigates into a value whose declared kind does not fit the path, such as reading a field from something that is not an object, or applying an index or filter to something that is not a list.	Active
TYPE_VARIABLE_NOT_IN_SCOPE	This formula uses a name that nothing defines at this point in the ruleset. A loop's item or index name can only be used inside the loop that declares it.	Active
TYPE_ASSERTION_REDUNDANT	This step declares an asserted type where the engine can already work the type out on its own.	Active
TYPE_ASSERTION_NOT_PERMITTED	This step's asserted type sits at a position whose type is unknown only because an earlier step in this ruleset already failed.	Active
SCHEMA_UNSUPPORTED_CONSTRUCT	A schema this ruleset declares - its input or output schema, a dataset's row schema or a step's asserted output type - uses a construct that rule-line's type system does not support.	Active
SCHEMA_REQUIRED_UNDECLARED_PROPERTY	This ruleset's declared schema requires a field that the same object does not declare as a property.	Active
SCHEMA_TOO_DEEPLY_NESTED	A schema this ruleset declares - its input or output schema, a dataset's row schema or a step's asserted output type - nests more deeply than rule-line can process.	Active
SCHEMA_TOO_LARGE	A schema this ruleset declares - its input or output schema, a dataset's row schema or a step's asserted output type - declares more schema elements in total than rule-line can process.	Active
STEP_ASSERTED_OUTPUT_MISMATCH	A step's real output did not match the shape this ruleset declared for it.	Active
GRAPH_STEP_REFERENCE_CYCLE	Two or more steps in this ruleset depend on each other in a loop, so no order can produce their values.	Active
GRAPH_STEP_REFERENCE_NOT_GUARANTEED	A formula references a step that is not guaranteed to have produced a value at that position.	Active
GRAPH_UNREACHABLE_STEP	This ruleset contains a step that no path through the ruleset can reach.	Active
GRAPH_OUTPUT_STEP_MISSING	A branch or foreach step names an output step that one of its bodies does not contain.	Active
STEP_GRAPH_DEPTH_EXCEEDED	This ruleset nests steps inside steps more deeply than rule-line allows.	Active
STEP_GRAPH_TOO_MANY_STEPS	This ruleset declares more total steps than rule-line allows for one document.	Active
GRAPH_MULTIPLE_ROOT_STEPS_UNSUPPORTED	This ruleset has more than one top-level step. A ruleset runs from exactly one top-level step in this version of rule-line; place the steps inside a single top-level step.	Active
DECIMAL_NOT_REPRESENTABLE	A decimal in this run's declared output cannot be represented exactly as a plain JSON number, so the run was refused rather than rounded.	Active
RESULT_OUTPUT_SCHEMA_MISMATCH	This run's output does not match the shape the ruleset declared for it.	Active
RESULT_OUTPUT_STEP_MISSING	The step this ruleset names as its output produced no value when the run reached it.	Active
RESULT_OUTPUT_SELECTOR_INVALID	This ruleset names an output step that is not a step at the top level of the ruleset.	Active
RESULT_MULTIPLE_ROOT_STEPS_UNSUPPORTED	This ruleset is valid, but running a ruleset with more than one top-level step is not yet supported by this version of rule-line.	Active
ENVELOPE_WALL_CLOCK_EXCEEDED	A run of this ruleset took longer to execute than its declared wall-clock limit allows.	Active
ENVELOPE_STEP_COUNT_EXCEEDED	A run of this ruleset executed more steps than its declared step-count limit allows.	Active
ENVELOPE_MEMORY_EXCEEDED	A run of this ruleset used more memory than its declared memory limit allows.	Active
ENVELOPE_FOREACH_EXCEEDED	A foreach step in this ruleset iterated more times than its declared per-loop iteration limit allows.	Active
ENVELOPE_MODEL_CALLS_EXCEEDED	A run of this ruleset made more AI model calls than its declared limit allows.	Active
STEP_EXECUTION_FAILED	A step's own evaluation failed, and the step declares this as a hard error rather than a recoverable one.	Active
STEP_SOFT_FALLBACK_APPLIED	A step's declared fallback value was used in place of a value it could not compute. The run continues, and the use of the fallback value is recorded in the run's trace.	Reserved — not yet emitted
STEP_FOREACH_COLLECTION_NOT_ARRAY	The collection a foreach step iterates did not evaluate to a list.	Active
STEP_BRANCH_NO_CASE_MATCHED	No case of a branch step matched the submitted data, and the step declares no fallback list.	Active
STEP_FOREACH_VAR_SHADOWS_RESERVED_ROOT	A loop variable in this ruleset is named the same as a reserved data-access root, so the variable can never be read. Reported as a warning; the ruleset stays valid.	Active
EXPR_CONTEXT_TOO_DEEPLY_NESTED	The data submitted for this run is nested more deeply than rule-line can process.	Active
XML_PARSE_FAILED	This input is not well-formed XML.	Active
XML_UNSUPPORTED_CONSTRUCT	This document uses a construct rule-line refuses to process, such as a document type declaration, an entity declaration or an inclusion directive.	Active
XML_MIXED_CONTENT_REFUSED	The declared schema allows an element to hold both text and child elements, which this conversion does not support, so the conversion was refused.	Active
XML_NAMESPACE_PREFIX_UNDECLARED	This document uses a namespace that the step does not give a prefix to.	Active
XML_SCHEMA_ARTIFACT_UNAVAILABLE	The exact schema version this step names is not available to the run, or the steps that use one schema declare different namespace prefixes (conversion.namespaces). Make the version available, or give every step on that schema the same prefixes.	Active
XML_VALIDATION_ARTIFACT_UNENFORCED	This step names a validation artifact that rule-line cannot apply, so the step is refused rather than accepted with the artifact ignored.	Active
DOCUMENT_INPUTS_INVALID_ROLE_NAME	This document input role name is not accepted. Use an ASCII identifier of 1 to 64 characters starting with a letter, and avoid the reserved names 'input', 'steps', and 'documents'.	Active
DOCUMENT_INPUTS_INVALID_CARDINALITY	This document input role declares a minimum or maximum document count that rule-line does not accept — both must be whole numbers between 0 and 20, and the maximum must not be less than the minimum.	Active
DOCUMENT_INPUTS_INVALID_MEDIA_TYPE	This document input role declares an accepted media type that rule-line does not support, declares no accepted media types at all, or repeats the same media type more than once.	Active
DOCUMENT_INPUTS_SUBMISSION_MISMATCH	The documents submitted with this job do not match the document input roles this ruleset declares — a role it does not declare, fewer documents than a role's declared minimum or more than its declared maximum, or a media type a role does not accept.	Active
DOCUMENT_PARSE_FAILED	A document accepted at submission could not be read again with unchanged, valid content when this run started — the run has finished without producing a result.	Active
ENGINE_TASK_ABORTED	This run's engine task did not return a result — the run was aborted after exceeding its execution time backstop, or the worker running it crashed. The run has finished without producing a result, and its trace may be missing rows.	Active
STEP_FOREACH_VAR_RESERVES_DOCUMENTS_ROOT	A foreach step's loop variable is named 'documents', which is a reserved data-access root and can never be bound as a variable. Rename the loop variable; this ruleset cannot be saved until it is.	Active
TEST_RUN_MUTATION_FORBIDDEN	This operation performs a business-data mutation, which is forbidden while running in TEST mode.	Reserved — not yet emitted
DATASET_NAME_INVALID	A dataset name is not accepted. Use an ASCII identifier of 1 to 64 characters starting with a letter, made of letters, digits and underscores.	Active
DATASET_COUNT_EXCEEDED	This ruleset declares more datasets than rule-line allows. A ruleset may declare at most 20 datasets.	Active
DATASET_SCHEMA_ROOT_NOT_OBJECT	A dataset schema describes one row, so it must be an object at the top level. Declare the row as an object with its fields as properties.	Active
DATASET_KEY_NOT_SCALAR_LEAF	A key field of this dataset does not name a single value in the row. A key must point to a text or number field, not to a missing field, a group of fields or a list.	Active
DATASET_KEY_TYPE_UNSUPPORTED	A key field of this dataset has a type that cannot identify a row. Every key field must be required, must not allow an empty value, and must be text or a whole number.	Active
DATASET_INDEX_NOT_SCALAR_LEAF	An indexed field of this dataset does not name a single value in the row. An index must point to a text, number or yes/no field, not to a missing field, a group of fields or a list.	Active
DATASET_INDEX_TYPE_UNSUPPORTED	An indexed field of this dataset mixes value kinds, such as text and number. An indexed field must hold one kind of value, optionally allowing an empty value.	Active
DATASET_INDEX_DUPLICATE	This dataset names the same field more than once among its key fields and indexes. Key fields are indexed automatically, so list each field only once.	Active
DATASET_SLOT_BUDGET_EXCEEDED	This dataset keys or indexes more fields of one kind than rule-line can store for fast lookup. At most 6 text, 6 number and 2 yes/no fields may be key or indexed fields together.	Active
LOOKUP_DATASET_VERSION_INVALID	A lookup step pins a dataset version that is not a valid version number. Use a whole number of up to 10 digits that does not start with zero.	Active
LOOKUP_NO_MATCH	A lookup step that expects exactly one row found none. The values it searched for are recorded in the run trace (a very large search may be shortened there).	Active
LOOKUP_AMBIGUOUS_MATCH	A lookup step that expects exactly one row found several. The search values and the identifying keys of the rows it found (a long list of rows, or a very large search, may be cut short) are recorded in the run trace.	Active
DATASET_NO_CONFORMING_VERSION	No stored version of a dataset matches the schema this ruleset now declares for it, so the ruleset cannot be activated. Push the data again, or restore the earlier declaration.	Active
DATASET_VERSION_UNAVAILABLE	The dataset version this run needs is no longer available, so the run could not read its reference data.	Active
DATASET_ROWS_NOT_ARRAY	The data sent for this dataset is not a list of rows. Send the rows as a list, one entry per row.	Active
DATASET_ROW_NOT_OBJECT	A row of this dataset is not an object. Each row must be an object with one entry per field.	Active
DATASET_ROW_MISSING_REQUIRED_FIELD	A row of this dataset leaves out a field the dataset requires. Add the field to the row.	Active
DATASET_ROW_UNKNOWN_FIELD	A row of this dataset carries a field the dataset does not declare. Remove the field from the row, or add it to the dataset declaration.	Active
DATASET_ROW_WRONG_TYPE	A field of a row holds the wrong kind of value, such as text where a number is declared, or an empty value where one is not allowed.	Active
DATASET_ROW_NOT_INTEGER	A field of a row holds a number that cannot be stored exactly: a value with decimals in a field declared as an integer, or a whole number too large to be stored exactly.	Active
DATASET_ROW_NOT_IN_ENUM	A field of a row holds a value that is not one of the values the dataset allows for that field.	Active
DATASET_ROW_UNSUPPORTED_CHARACTER	A text field of a row contains a character that cannot be stored, such as a null character or half of a paired symbol. Remove or replace the character.	Active
DATASET_ROW_DUPLICATE_KEY	Two entries of this push share the same key. A key identifies exactly one row, so keep only one of the entries.	Active
DATASET_DELETE_KEY_INVALID	An entry to delete does not consist of exactly the key fields of the dataset with valid values. Name only the key fields, each with a value of the declared kind.	Active
LOOKUP_DATASET_UNDECLARED	A lookup step names a dataset that this ruleset does not declare. While the ruleset is a draft this is a warning and the lookup result stays untyped; the ruleset cannot be activated until the dataset is declared.	Active
LOOKUP_FIELD_UNDECLARED	A lookup step matches on a field that the dataset declaration does not contain. While the ruleset is a draft this is a warning; the ruleset cannot be activated until the field is declared or the match is changed.	Active
LOOKUP_FIELD_NOT_INDEXED	A lookup step matches on a field that is neither a key field nor an indexed field of the dataset, or that names a group of fields or a list. Match only on key fields or indexed fields, or add the field to the dataset indexes.	Active
LOOKUP_OPERATOR_TYPE_MISMATCH	A lookup match compares a field with an operator or a value its type does not allow, such as contains on a number, approxEqual on text, or a text value against a number field. Use an operator and a value that fit the field type.	Active
LOOKUP_FLAG_NOT_STRING_FIELD	A lookup match applies a text normalization flag, such as trim or caseInsensitive, to a field that is not text. Remove the flag or match on a text field.	Active
LOOKUP_ORDER_BY_FIELD_NOT_INDEXED	A lookup step orders its rows by a field that is neither a key field nor an indexed field of the dataset. Order only by key fields or indexed fields, or add the field to the dataset indexes.	Active
DATASET_DECLARATION_TOO_DEEP	The row schema of this dataset nests groups inside groups more deeply than rule-line can store, so the dataset cannot be saved. Flatten the schema to at most 32 levels.	Active
EXTRACT_MODEL_ALIAS_UNKNOWN	An extract step names a model alias that is not in the model catalog. Choose an alias the catalog lists before activating the ruleset.	Active
EXTRACT_DOCUMENT_MEDIA_TYPE_UNREADABLE	The document role an extract step reads accepts file types that extraction cannot read. Extraction reads PDF, PNG, JPEG, TIFF, plain text, HTML and email (.eml); remove the other types from the role.	Active
EXTRACT_DOCUMENT_XML_ONLY	The document role an extract step reads accepts only XML, which an extract step does not read. Use an XML transform step for structured documents, or branch on the media type.	Active
EXTRACT_READING_NO_EFFECT	An extract step sets a reading mode, but the document role it reads accepts only text, HTML or email documents, which are read as text whatever the mode. The setting has no effect.	Active
EXTRACT_TARGET_SCHEMA_NOT_STRICT	Part of the target schema of an extract step cannot be enforced strictly by the model, so the model could return a value that breaks it. Simplify that part of the schema or make the field required.	Active
EXTRACT_TARGET_SCHEMA_RESERVED_NAME	A field in the target schema of an extract step uses a name the platform reserves for its own use. Rename the field.	Active
EXTRACT_RETRIES_EXCEED_MODEL_CALLS	The corrective retries an extract step allows could need more model calls than the ruleset allows a run to make. Lower the retries or raise the model-call limit of the ruleset.	Active
EXTRACT_PROMPT_EMPTY	An extract step uses a custom task instruction, but the instruction is empty. Write an instruction, or switch back to the generated prompt.	Active
EXTRACT_PLAUSIBILITY_NAME_DUPLICATE	Two plausibility checks on one extract step have the same name. Give each check its own name so a result can say which check failed.	Active
EXTRACT_SCHEMA_MISMATCH	The answer of the model did not match the target schema of the extract step, even after the allowed corrective retries, so it was not accepted as the output of the step.	Active
EXTRACT_INJECTION_SUSPECTED	The document contains text that looks like an attempt to instruct the model, so extraction stopped. The flags in the extraction result of the step say what was found.	Active
EXTRACT_CALL_TIMEOUT	A model call of an extract step took longer than the allowed time and was stopped.	Active
EXTRACT_DOCUMENT_XML	The document handed to an extract step is XML, which an extract step does not read. Use an XML transform step for structured documents.	Active
EXTRACT_DOCUMENT_UNREADABLE	The document handed to an extract step could not be read for extraction, so extraction did not run. The failure's detail code names the cause: the document holds no readable text, is a TIFF with several pages, is a file type the OCR model does not accept, came back from OCR with a missing page, has an empty page under a forced text-layer reading, has more pages than the platform reads, or has no reading on file.	Active
EXTRACT_UNREADABLE_NO_READABLE_CHARACTERS	Detail of EXTRACT_DOCUMENT_UNREADABLE: no page of the document holds readable text, so there is nothing for an extract step to read.	Active
EXTRACT_UNREADABLE_TIFF_MULTI_PAGE	Detail of EXTRACT_DOCUMENT_UNREADABLE: the document is a TIFF with more than one page, and an extract step reads only single-page TIFF files.	Active
EXTRACT_UNREADABLE_OCR_MEDIA_TYPE	Detail of EXTRACT_DOCUMENT_UNREADABLE: the document has to be read by OCR, but the OCR model of the platform accepts images only and this file is not an image.	Active
EXTRACT_UNREADABLE_OCR_PAGES_MISSING	Detail of EXTRACT_DOCUMENT_UNREADABLE: the OCR model did not return a result for every page of the document, and an extract step never works on part of a document.	Active
EXTRACT_UNREADABLE_TEXT_LAYER_EMPTY	Detail of EXTRACT_DOCUMENT_UNREADABLE: the step's reading mode is set to the text layer, and at least one page of the document has no text layer.	Active
EXTRACT_UNREADABLE_NO_READING	Detail of EXTRACT_DOCUMENT_UNREADABLE: the platform holds no reading of the document, so there is nothing for an extract step to read.	Active
EXTRACT_UNREADABLE_PAGES_TRUNCATED	Detail of EXTRACT_DOCUMENT_UNREADABLE: the document has more pages than the platform reads, and an extract step never works on part of a document.	Active
EXTRACT_MODEL_CANNOT_ENFORCE_SCHEMA	The model alias chosen for an extract step cannot enforce a response schema, so the step cannot extract with it. The analysis of a ruleset reports this before any run, and a run reports it when the step starts, before any model call. Choose a different model alias.	Active
EXTRACT_PROVIDER_REFUSED	The model provider refused to process the document handed to an extract step, or its answer was larger than the platform accepts, so extraction did not run.	Active
EXTRACT_OCR_UNAVAILABLE	The document handed to an extract step needs optical character recognition (OCR) to be read, but no OCR model is configured, so extraction did not run.	Active
ENVELOPE_PAGES_EXCEEDED	A run of this ruleset would have sent more document pages to extraction than its declared page limit allows, so the step stopped before calling the model.	Active