diffusers-workflow 0.4.0b5__tar.gz → 0.4.0b6__tar.gz

{diffusers_workflow-0.4.0b5/diffusers_workflow.egg-info → diffusers_workflow-0.4.0b6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffusers-workflow
-Version: 0.4.0b5
+Version: 0.4.0b6
 Summary: Declarative workflow engine and web UI for Hugging Face Diffusers
 Author: dkackman
 License: Apache-2.0
@@ -19,7 +19,7 @@ License-File: LICENSE
 Requires-Dist: torch>=2.13.0
 Requires-Dist: torchvision>=0.28.0
 Requires-Dist: diffusers>=0.40.0
-Requires-Dist: transformers>=5.16.1
+Requires-Dist: transformers!=5.17.0,>=5.16.1
 Requires-Dist: accelerate>=1.14.0
 Requires-Dist: peft>=0.20.0
 Requires-Dist: safetensors>=0.8.0
@@ -36,6 +36,7 @@ Requires-Dist: bitsandbytes>=0.50.2; sys_platform == "linux"
 Requires-Dist: controlnet-aux>=0.0.10
 Requires-Dist: opencv-python-headless>=5.0.0.93
 Requires-Dist: kornia>=0.8.3
+Requires-Dist: kernels>=0.17.0
 Requires-Dist: spandrel>=0.4.2
 Requires-Dist: facexlib>=0.3.0
 Requires-Dist: numpy>=2.5.2

{diffusers_workflow-0.4.0b5 → diffusers_workflow-0.4.0b6/diffusers_workflow.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffusers-workflow
-Version: 0.4.0b5
+Version: 0.4.0b6
 Summary: Declarative workflow engine and web UI for Hugging Face Diffusers
 Author: dkackman
 License: Apache-2.0
@@ -19,7 +19,7 @@ License-File: LICENSE
 Requires-Dist: torch>=2.13.0
 Requires-Dist: torchvision>=0.28.0
 Requires-Dist: diffusers>=0.40.0
-Requires-Dist: transformers>=5.16.1
+Requires-Dist: transformers!=5.17.0,>=5.16.1
 Requires-Dist: accelerate>=1.14.0
 Requires-Dist: peft>=0.20.0
 Requires-Dist: safetensors>=0.8.0
@@ -36,6 +36,7 @@ Requires-Dist: bitsandbytes>=0.50.2; sys_platform == "linux"
 Requires-Dist: controlnet-aux>=0.0.10
 Requires-Dist: opencv-python-headless>=5.0.0.93
 Requires-Dist: kornia>=0.8.3
+Requires-Dist: kernels>=0.17.0
 Requires-Dist: spandrel>=0.4.2
 Requires-Dist: facexlib>=0.3.0
 Requires-Dist: numpy>=2.5.2

{diffusers_workflow-0.4.0b5 → diffusers_workflow-0.4.0b6}/diffusers_workflow.egg-info/SOURCES.txt

@@ -9,15 +9,19 @@ diffusers_workflow.egg-info/entry_points.txt
 diffusers_workflow.egg-info/requires.txt
 diffusers_workflow.egg-info/top_level.txt
 dw/__init__.py
+dw/adapter_compatibility.py
 dw/arguments.py
 dw/assets.py
 dw/cache_blocks.json
 dw/cache_blocks.py
+dw/content_types.py
+dw/elision.py
 dw/events.py
 dw/for_each.py
 dw/host_memory.py
 dw/hub_cache.py
 dw/introspection.py
+dw/kernel_availability.py
 dw/locations.py
 dw/log_setup.py
 dw/media_info.py
@@ -28,6 +32,7 @@ dw/prompt_weighting.py
 dw/prompts.py
 dw/realize.py
 dw/reference_limits.py
+dw/reference_names.py
 dw/repl.py
 dw/repl_commands.py
 dw/repl_worker.py
@@ -47,6 +52,7 @@ dw/teacache_models.json
 dw/test.py
 dw/type_helpers.py
 dw/validate.py
+dw/variable_constraints.py
 dw/variables.py
 dw/worker.py
 dw/workflow.py
@@ -87,6 +93,7 @@ dw/server/guides.py
 dw/server/jobs.py
 dw/server/mcp_mount.py
 dw/server/netinfo.py
+dw/server/observed_cost.py
 dw/server/updater.py
 dw/server/ui/index.html
 dw/server/ui/assets/abap-08VXUWAP.js
@@ -103,45 +110,45 @@ dw/server/ui/assets/csharp-BKxtCVv1.js
 dw/server/ui/assets/csp-bTuwJoIa.js
 dw/server/ui/assets/css-DIMkf-bt.js
 dw/server/ui/assets/css.worker-B3ciXF_0.js
-dw/server/ui/assets/cssMode-BQ03gENP.js
+dw/server/ui/assets/cssMode-1t4O7A8b.js
 dw/server/ui/assets/cypher-CVaqCwHa.js
 dw/server/ui/assets/dart-onAF5SnQ.js
 dw/server/ui/assets/dockerfile-DZFCIeNp.js
 dw/server/ui/assets/ecl-D05T4iGw.js
 dw/server/ui/assets/editor-jjEx9u7D.css
-dw/server/ui/assets/editor.api-CzvLX8mb.js
+dw/server/ui/assets/editor.api-C2mqCHa9.js
 dw/server/ui/assets/editor.worker-q-txB4vs.js
 dw/server/ui/assets/elixir-6RTg0lbw.js
 dw/server/ui/assets/flow9-C5_-GSwl.js
-dw/server/ui/assets/freemarker2-C8c7lwsP.js
+dw/server/ui/assets/freemarker2-Cb6RGT42.js
 dw/server/ui/assets/fsharp-C8Ef5oNN.js
 dw/server/ui/assets/go-C-y9NEjX.js
 dw/server/ui/assets/graphql-fmXr3nnJ.js
-dw/server/ui/assets/handlebars-7APo_5hD.js
+dw/server/ui/assets/handlebars-LzK4_SEF.js
 dw/server/ui/assets/hcl-CpzslTdj.js
-dw/server/ui/assets/html-JgrkBemQ.js
+dw/server/ui/assets/html-Biz_qd50.js
 dw/server/ui/assets/html.worker-C93Ht9o9.js
-dw/server/ui/assets/htmlMode-BH9VsgGh.js
-dw/server/ui/assets/index-B1DfbIuu.js
-dw/server/ui/assets/index-CRVAbEJ9.css
+dw/server/ui/assets/htmlMode-CP4Kjl0x.js
+dw/server/ui/assets/index-Bt8BlCkf.js
+dw/server/ui/assets/index-D9tEUhx5.css
 dw/server/ui/assets/ini-sBoK_t0W.js
 dw/server/ui/assets/java-BEtHBSE6.js
-dw/server/ui/assets/javascript-BrFJ1H6j.js
+dw/server/ui/assets/javascript-DX4-t0BE.js
 dw/server/ui/assets/json.worker-B2V3pomh.js
-dw/server/ui/assets/jsonMode-C10gRe-2.js
+dw/server/ui/assets/jsonMode-hZa_qOEa.js
 dw/server/ui/assets/julia-Bri6UV-V.js
 dw/server/ui/assets/kotlin-BOotOW0E.js
 dw/server/ui/assets/less-B9JPFI3C.js
 dw/server/ui/assets/lexon-CfSJPG6W.js
-dw/server/ui/assets/liquid-CjfGWTAZ.js
-dw/server/ui/assets/lspLanguageFeatures-O-lB5SDv.js
+dw/server/ui/assets/liquid-BRkTuJ_A.js
+dw/server/ui/assets/lspLanguageFeatures-Dvt5tod9.js
 dw/server/ui/assets/lua-CsQS60Ue.js
 dw/server/ui/assets/m3-D-oSqn_W.js
 dw/server/ui/assets/markdown-Cimd5fb3.js
-dw/server/ui/assets/mdx-D7b-hMTY.js
+dw/server/ui/assets/mdx-Bdet7qev.js
 dw/server/ui/assets/mips-CIPQ_RoX.js
 dw/server/ui/assets/monaco--ixms01u.css
-dw/server/ui/assets/monaco-DBnxML20.js
+dw/server/ui/assets/monaco-QRyCMGAv.js
 dw/server/ui/assets/msdax-DauUninz.js
 dw/server/ui/assets/mysql-SOo6toE5.js
 dw/server/ui/assets/objective-c-FvmIjYaQ.js
@@ -156,10 +163,10 @@ dw/server/ui/assets/powerquery-D3hlyOfw.js
 dw/server/ui/assets/powershell-DmHpPYUd.js
 dw/server/ui/assets/protobuf-C531GsRP.js
 dw/server/ui/assets/pug-Z5eAx3Zn.js
-dw/server/ui/assets/python-NfJjzkPL.js
+dw/server/ui/assets/python-DO0uJyUr.js
 dw/server/ui/assets/qsharp-DkqhCAOL.js
 dw/server/ui/assets/r-BwWrilGY.js
-dw/server/ui/assets/razor-CgQtwMBj.js
+dw/server/ui/assets/razor-CkfVe6oz.js
 dw/server/ui/assets/redis-ClamHrr6.js
 dw/server/ui/assets/redshift-DT7zqm-g.js
 dw/server/ui/assets/restructuredtext-BYgofb2h.js
@@ -179,15 +186,15 @@ dw/server/ui/assets/swift-Bxkupp3x.js
 dw/server/ui/assets/systemverilog-Bz4Y3fRF.js
 dw/server/ui/assets/tcl-DISqw1ZD.js
 dw/server/ui/assets/ts.worker-D7T1-Ig5.js
-dw/server/ui/assets/tsMode-BuMW6Qar.js
+dw/server/ui/assets/tsMode-z26QXNyE.js
 dw/server/ui/assets/twig-De2hgUGE.js
-dw/server/ui/assets/typescript-Bfmxt754.js
+dw/server/ui/assets/typescript-DIdNc_uc.js
 dw/server/ui/assets/typespec-B8J7ngcE.js
 dw/server/ui/assets/vb-DV3o63ZY.js
 dw/server/ui/assets/wgsl-DpFanUEy.js
-dw/server/ui/assets/workers-BSu9lD5s.js
-dw/server/ui/assets/xml-BJ2IwGRC.js
-dw/server/ui/assets/yaml-Cq85tEmC.js
+dw/server/ui/assets/workers-Ce8CDVu5.js
+dw/server/ui/assets/xml-Djj2U0Rl.js
+dw/server/ui/assets/yaml-CePLO71e.js
 dw/tasks/audio_utils.py
 dw/tasks/background_remover.py
 dw/tasks/borders.py
@@ -247,6 +254,7 @@ tests/test_compose_text.py
 tests/test_concat_videos.py
 tests/test_config_objects.py
 tests/test_configuration_schema.py
+tests/test_content_types.py
 tests/test_depth_estimator.py
 tests/test_device.py
 tests/test_device_helpers.py
@@ -255,11 +263,14 @@ tests/test_dialogue_short_voices.py
 tests/test_diffusion_upscale.py
 tests/test_dissolve_videos.py
 tests/test_docs_links.py
+tests/test_elision.py
 tests/test_events.py
 tests/test_examples.py
 tests/test_for_each.py
 tests/test_gather.py
+tests/test_h3_adapters.py
 tests/test_h3_context_ir.py
+tests/test_h3_schedule.py
 tests/test_hf_placement.py
 tests/test_host_memory.py
 tests/test_hub_cache.py
@@ -269,11 +280,16 @@ tests/test_integration.py
 tests/test_interpolate_frames.py
 tests/test_introspection.py
 tests/test_job_progress.py
+tests/test_job_warnings.py
 tests/test_jobs_listing.py
 tests/test_jobs_reused.py
+tests/test_kernel_availability.py
 tests/test_library_sources.py
 tests/test_list_images.py
 tests/test_locations.py
+tests/test_lora_alpha.py
+tests/test_ltx2_diffusion_decode.py
+tests/test_ltx2_ic_loras.py
 tests/test_ltx_prompt_library.py
 tests/test_mcp_assets.py
 tests/test_mcp_authoring.py
@@ -293,9 +309,11 @@ tests/test_mix_audio.py
 tests/test_model_cache.py
 tests/test_modular_pipeline.py
 tests/test_modular_progress.py
+tests/test_observed_cost.py
 tests/test_offload_placement.py
 tests/test_output_references.py
 tests/test_pair_audio.py
+tests/test_pair_audio_fit.py
 tests/test_phase_events.py
 tests/test_pipeline_caching.py
 tests/test_pipeline_components.py
@@ -309,6 +327,7 @@ tests/test_prompts.py
 tests/test_realize.py
 tests/test_recenter_crop.py
 tests/test_reference_limits.py
+tests/test_reference_names.py
 tests/test_repl_commands.py
 tests/test_repl_ergonomics.py
 tests/test_repl_hierarchical.py
@@ -341,6 +360,7 @@ tests/test_subfolders.py
 tests/test_task.py
 tests/test_task_discovery.py
 tests/test_task_domains.py
+tests/test_task_signature_errors.py
 tests/test_teacache.py
 tests/test_template_subfolders.py
 tests/test_tensor_image.py
@@ -349,6 +369,7 @@ tests/test_text_sections.py
 tests/test_type_helpers.py
 tests/test_updater.py
 tests/test_validate_arguments.py
+tests/test_variable_constraints.py
 tests/test_variables.py
 tests/test_video_utils.py
 tests/test_worker.py

{diffusers_workflow-0.4.0b5 → diffusers_workflow-0.4.0b6}/diffusers_workflow.egg-info/requires.txt

@@ -1,7 +1,7 @@
 torch>=2.13.0
 torchvision>=0.28.0
 diffusers>=0.40.0
-transformers>=5.16.1
+transformers!=5.17.0,>=5.16.1
 accelerate>=1.14.0
 peft>=0.20.0
 safetensors>=0.8.0
@@ -17,6 +17,7 @@ gguf>=0.19.0
 controlnet-aux>=0.0.10
 opencv-python-headless>=5.0.0.93
 kornia>=0.8.3
+kernels>=0.17.0
 spandrel>=0.4.2
 facexlib>=0.3.0
 numpy>=2.5.2

diffusers_workflow-0.4.0b6/dw/adapter_compatibility.py

@@ -0,0 +1,226 @@
+"""The LoRA a step loads has to be trained for the path that step runs.
+
+MiniMax-H3 ships one repository holding two checkpoint partitions -
+`transformer/` for the `t2va` and `fl2va` workflows, `transformer_ref/` for
+`ref2va` - and a separate turbo adapter trained against each. Handing the
+FL2VA adapter to a `ref2va` step is the one H3 misconfiguration that never
+shows up as a failure: a `ref2va` step holds `transformer_ref` alone, so
+diffusers loads whatever it is given straight onto it, the run succeeds, and
+the only symptom is identity retention that quietly is not as good as it
+should be. Four templates carried exactly that mistake and nothing ever
+complained (#149, #155).
+
+Everything a pipeline itself would refuse is refused by the pipeline. This
+is the other kind: a request the pipeline accepts and answers wrongly, which
+costs a whole render to discover and cannot be seen in the output. So it is
+checked where it is free - `validation_errors`, so `POST /api/validate` and
+the pre-queue check both refuse it, at the JSON path the value sits at.
+
+What is *not* written here: the adapter file names. `lora_weight_name` is a
+free string and a future reference-trained checkpoint cannot be predicted, so
+an unrecognised name is a warning naming the rule rather than a refusal - the
+default inverts from "anything passes" to "the one documented mistake is
+caught", and someone testing `my-new-ref-lora.safetensors` still gets
+through. What *is* written here is a vendor naming convention (`ref2v` /
+`fl2v` in the file name) that no diffusers symbol declares; the workflow
+names and the partition each one denoises against are diffusers' own, and
+`tests/test_h3_adapters.py` pins them to it.
+"""
+
+import logging
+
+from .for_each import MEMBER_SEPARATOR, render_path
+
+logger = logging.getLogger("dw")
+
+# The `workflow=` values MiniMaxH3Blocks._workflow_map declares, split by the
+# transformer partition each one denoises against - a `ref2va` step loads
+# `transformer_ref` and nothing else, every other H3 workflow loads
+# `transformer`. A workflow name not listed here is not checked
+REFERENCE_WORKFLOWS = frozenset({"ref2va"})
+KEYFRAME_WORKFLOWS = frozenset({"t2va", "fl2va"})
+H3_WORKFLOWS = REFERENCE_WORKFLOWS | KEYFRAME_WORKFLOWS
+
+# The token MiniMax puts in an adapter's file name to say which partition it
+# was trained against. `ref2v` first: it is the longer match and a name
+# carrying both is ambiguous rather than a keyframe adapter
+REFERENCE_TOKEN = "ref2v"
+KEYFRAME_TOKEN = "fl2v"
+
+LORAS_KEY = "loras"
+WEIGHT_NAME_KEY = "weight_name"
+WORKFLOW_KEY = "workflow"
+FROM_PRETRAINED_KEY = "from_pretrained_arguments"
+VARIABLE_PREFIX = "variable:"
+# Values another pass resolves; one still spelled out here is not this
+# pass's complaint
+_UNRESOLVED_PREFIXES = ("variable:", "item:", "previous_result:", "gather:")
+
+
+def _trained_for(weight_name):
+    """Which partition an adapter's file name says it was trained against -
+    'reference', 'keyframe', or None when the name says neither."""
+    lowered = weight_name.lower()
+    if REFERENCE_TOKEN in lowered:
+        return "reference"
+    if KEYFRAME_TOKEN in lowered:
+        return "keyframe"
+    return None
+
+
+def _problem(workflow, weight_name):
+    """(severity, message) for one adapter on one step, or None when the
+    pairing is right."""
+    trained = _trained_for(weight_name)
+    wants = "reference" if workflow in REFERENCE_WORKFLOWS else "keyframe"
+    if trained == wants:
+        return None
+    if trained is None:
+        return (
+            "warning",
+            f"'{weight_name}' names neither '{REFERENCE_TOKEN}' nor "
+            f"'{KEYFRAME_TOKEN}', so it cannot be checked against this "
+            f"step's '{workflow}' workflow. A '{workflow}' step denoises "
+            f"against the "
+            f"'{'transformer_ref' if wants == 'reference' else 'transformer'}' "
+            f"partition, and an adapter trained against the other one loads "
+            f"onto it without error and only degrades the result",
+        )
+    return (
+        "error",
+        f"'{weight_name}' is trained against the "
+        f"'{'transformer_ref' if trained == 'reference' else 'transformer'}' "
+        f"partition and this step runs the '{workflow}' workflow, which "
+        f"denoises against "
+        f"'{'transformer_ref' if wants == 'reference' else 'transformer'}'. "
+        f"MiniMax-H3 loads it anyway and the run succeeds - the only symptom "
+        f"is a worse result - so it is refused here. Use an adapter trained "
+        f"for this workflow (its name carries "
+        f"'{REFERENCE_TOKEN if wants == 'reference' else KEYFRAME_TOKEN}')",
+    )
+
+
+def _lora_problems(steps, source_indices, written=None, supplied=()):
+    """Every adapter/workflow mismatch in an expanded step list, as
+    (severity, path, message).
+
+    `written` is the definition as the author wrote it and `supplied` the
+    argument names the caller passed: when the value arrived through a
+    variable the caller set, the path reported is `arguments.<name>`, where
+    they wrote it, rather than the step it landed in.
+    """
+    written_steps = (written or {}).get("steps")
+    found = []
+    for index, step in enumerate(steps):
+        if not isinstance(step, dict):
+            continue
+        pipeline = step.get("pipeline")
+        if not isinstance(pipeline, dict):
+            continue
+        source = (
+            source_indices[index]
+            if source_indices is not None and index < len(source_indices)
+            else index
+        )
+        from_pretrained = pipeline.get(FROM_PRETRAINED_KEY)
+        workflow = (
+            from_pretrained.get(WORKFLOW_KEY)
+            if isinstance(from_pretrained, dict)
+            else None
+        )
+        if workflow not in H3_WORKFLOWS:
+            continue
+        loras = pipeline.get(LORAS_KEY)
+        if not isinstance(loras, list):
+            continue
+        name = step.get("name")
+        where = (
+            f" in member '{name}'"
+            if isinstance(name, str) and MEMBER_SEPARATOR in name
+            else ""
+        )
+        for position, lora in enumerate(loras):
+            if not isinstance(lora, dict):
+                continue
+            weight_name = lora.get(WEIGHT_NAME_KEY)
+            if not isinstance(weight_name, str) or weight_name.startswith(
+                _UNRESOLVED_PREFIXES
+            ):
+                continue
+            problem = _problem(workflow, weight_name)
+            if problem is None:
+                continue
+            severity, message = problem
+            found.append(
+                (
+                    severity,
+                    _path_for(written_steps, source, position, supplied)
+                    or render_path(
+                        (
+                            "steps",
+                            source,
+                            "pipeline",
+                            LORAS_KEY,
+                            position,
+                            WEIGHT_NAME_KEY,
+                        )
+                    ),
+                    message + where,
+                )
+            )
+    return found
+
+
+def _path_for(written_steps, source, position, supplied):
+    """`arguments.<name>` when the adapter came from a variable the caller
+    set, else None - the value is reported where it was written."""
+    if not isinstance(written_steps, list) or source >= len(written_steps):
+        return None
+    step = written_steps[source]
+    pipeline = step.get("pipeline") if isinstance(step, dict) else None
+    loras = pipeline.get(LORAS_KEY) if isinstance(pipeline, dict) else None
+    if not isinstance(loras, list) or position >= len(loras):
+        return None
+    entry = loras[position]
+    reference = entry.get(WEIGHT_NAME_KEY) if isinstance(entry, dict) else None
+    if not isinstance(reference, str) or not reference.startswith(VARIABLE_PREFIX):
+        return None
+    variable = reference.removeprefix(VARIABLE_PREFIX)
+    return f"arguments.{variable}" if variable in (supplied or ()) else None
+
+
+def adapter_errors(workflow_definition, source_indices=None, written=None, supplied=()):
+    """Every adapter loaded onto the wrong checkpoint partition, as
+    [{path, message}]. The definition is the expanded, substituted one, so a
+    `for_each` member's own adapter is checked as it will run."""
+    return [
+        {"path": path, "message": message}
+        for severity, path, message in _lora_problems(
+            workflow_definition.get("steps") or [], source_indices, written, supplied
+        )
+        if severity == "error"
+    ]
+
+
+def adapter_warnings(
+    workflow_definition, source_indices=None, written=None, supplied=()
+):
+    """Every adapter whose name says nothing about what it was trained for,
+    as messages - valid, and said out loud because nothing at run time will."""
+    return [
+        f"{path}: {message}"
+        for severity, path, message in _lora_problems(
+            workflow_definition.get("steps") or [], source_indices, written, supplied
+        )
+        if severity == "warning"
+    ]
+
+
+def warn_adapters(workflow_definition):
+    """Say the unrecognised-adapter warning where whoever asked for the run
+    can read it - a run started from the CLI or a rerun never passed through
+    the validate route."""
+    from .events import emit_warning
+
+    for message in adapter_warnings(workflow_definition):
+        emit_warning(message, kind="adapter_unrecognized")

diffusers_workflow-0.4.0b6/dw/content_types.py

@@ -0,0 +1,105 @@
+"""A step's result 'content_type': the MIME type its writer is chosen by.
+
+`content_type: "video"` validated clean and then died deep inside a writer,
+in a traceback naming neither the field nor the value (#168) - the same
+shape as #162 in a different field. The writer's own dispatch matches on
+`content_type.startswith("video")`, so the bare word "video" took the video
+branch anyway; having no real MIME type it also had no extension to write,
+and the closest writer imageio could guess from an empty one was not a video
+writer at all.
+
+Audio and video each go through exactly one container this engine writes -
+`AUDIO_FORMATS` in `dw/result.py`, and the single `video/mp4` mux - so
+anything else in either family is refused here rather than accepted only to
+mismatch its writer later. image/*, text/* and *.json values stay
+permissive beyond the MIME-shape check: their writer dispatch is a generic
+prefix/suffix match (PIL's own format inference, a literal text or JSON
+write) with no narrower container to enforce a whitelist against.
+"""
+
+from .for_each import MEMBER_SEPARATOR, render_path
+from .result import AUDIO_FORMATS, MUXED_VIDEO_CONTENT_TYPE
+from .security import InvalidInputError, validate_content_type
+
+CONTENT_TYPE_KEY = "content_type"
+
+# Reference prefixes substitution resolves before this pass runs. One still
+# spelled out here is one nothing resolved, and that is the undeclared-
+# variable pass's complaint rather than a shape error
+_UNRESOLVED_PREFIXES = ("variable:", "item:")
+
+
+def content_type_fault(value):
+    """Why this result 'content_type' is invalid, or None.
+
+    Raises nothing - callers that already have an InvalidInputError-raising
+    check (validate_content_type) can call that directly; this is the
+    string-message form `content_type_errors` collects.
+    """
+    try:
+        validate_content_type(value)
+    except InvalidInputError as e:
+        return str(e)
+
+    main_type = value.split("/", 1)[0]
+    if main_type == "audio" and value not in AUDIO_FORMATS:
+        return (
+            f"Invalid content_type: {value!r} - audio can be written as "
+            f"{', '.join(sorted(AUDIO_FORMATS))}"
+        )
+    if main_type == "video" and value != MUXED_VIDEO_CONTENT_TYPE:
+        return (
+            f"Invalid content_type: {value!r} - video is only written as "
+            f"'{MUXED_VIDEO_CONTENT_TYPE}'"
+        )
+    return None
+
+
+def content_type_errors(workflow_definition, source_indices=None):
+    """Every result 'content_type' that no writer will accept, as
+    [{path, message}].
+
+    The definition handed here has already been substituted and expanded,
+    so every value in it is literal; a 'variable:' or 'item:' still spelled
+    out is left alone. `source_indices`, when given, is the source step
+    index of each step - a 'for_each' group turns one written step into
+    several, and the path an error carries has to be one the author can
+    find in the file they wrote; the member is named in the message.
+    """
+    steps = workflow_definition.get("steps")
+    if not isinstance(steps, list):
+        return []
+
+    errors = []
+    for index, step in enumerate(steps):
+        if not isinstance(step, dict):
+            continue
+        result = step.get("result")
+        if not isinstance(result, dict) or CONTENT_TYPE_KEY not in result:
+            continue
+        value = result[CONTENT_TYPE_KEY]
+        if isinstance(value, str) and value.startswith(_UNRESOLVED_PREFIXES):
+            continue
+        source = (
+            source_indices[index]
+            if source_indices is not None and index < len(source_indices)
+            else index
+        )
+        name = step.get("name")
+        where = (
+            f" in member '{name}'"
+            if isinstance(name, str) and MEMBER_SEPARATOR in name
+            else ""
+        )
+        fault = content_type_fault(value)
+        if fault is not None:
+            errors.append(
+                {
+                    "path": render_path(("steps", source, "result", CONTENT_TYPE_KEY)),
+                    "message": f"{fault}{where}",
+                }
+            )
+    return errors
+
+
+__all__ = ["content_type_errors", "content_type_fault"]

{diffusers_workflow-0.4.0b5 → diffusers_workflow-0.4.0b6}/dw/docs/MCP.md

@@ -226,7 +226,7 @@ when no single workflow covers it.
 | `get_health()` | — | Check that the server is alive, and which machine answered: `version`, `device`, whether the worker process is up, the job running now and the queue depth |
 | `get_server_info()` | — | What this installation can do and where it keeps things: `device` (the accelerator a run will use), `version`, the `workspace` this session is working in and the workflow/asset/output/prompt `directories` of *that* workspace, the bind address and port, whether a token is required, and whether MCP is mounted. Check the device before authoring - a CUDA-only choice (bitsandbytes, `torch.compile`, flash attention) is not available on an `mps` or `cpu` server |
 | `list_jobs(limit=20, status=None, workspace=None)` | optional `limit` (newest N), `status` (one state or a comma-separated set of `queued`, `running`, `succeeded`, `failed`, `cancelled`), `workspace` | List queued, running and recent jobs, **newest first**. Bounded by default: the unbounded listing was over a client's tool-result limit on a server with a few months of history, which made it a tool that could not be called at all. `total` says how many matched and `truncated`/`next` say so when the answer was cut - raise `limit` or narrow with `status`. Without `workspace`, a named workspace lists its own jobs and the default one lists every job the server holds |
-| `list_gallery(limit=50, subfolder=None, workspace=None)` | `limit`, `subfolder`, `workspace` | List generated output files, newest first. A name is `<workflow>/<run id>/<file>`, where `<file>` may sit in the subfolder the step chose (`final/episode.mp4`); each entry carries `folder` (the workflow) and `subfolder` (by convention `final` or `intermediate`, `''` when the step chose none, any path the workflow wrote otherwise), and `subfolder="final"` lists only deliverables. Each entry also carries a ready-made `url`, already scoped to the workspace that made it - a hand-built `/outputs/<name>` URL 404s for anything but the default workspace. `workspace` names the workspace for this one call without switching the session to it - the same pin `run_workflow` takes, so a job run into another workspace stays reachable from the session that queued it |
+| `list_gallery(limit=50, subfolder=None, only_orphans=False, workspace=None)` | `limit`, `subfolder`, `only_orphans`, `workspace` | List generated output files, newest first. A name is `<workflow>/<run id>/<file>`, where `<file>` may sit in the subfolder the step chose (`final/episode.mp4`); each entry carries `folder` (the workflow) and `subfolder` (by convention `final` or `intermediate`, `''` when the step chose none, any path the workflow wrote otherwise), and `subfolder="final"` lists only deliverables. Each entry also carries a ready-made `url`, already scoped to the workspace that made it - a hand-built `/outputs/<name>` URL 404s for anything but the default workspace. `only_orphans=True` inverts the call: instead of files, it returns run directories with no media anywhere under them (`runs`, each `{name, mtime}`) - a run whose output was deleted before `delete_output` could remove it by name, or one that failed before writing anything; `subfolder` does not apply in this mode, and `name` is exactly what `delete_output` accepts (#170). `workspace` names the workspace for this one call without switching the session to it - the same pin `run_workflow` takes, so a job run into another workspace stays reachable from the session that queued it |
 | `get_gallery_metadata(name, envelope=False, workspace=None)` | `name`, `workspace` | Get the metadata embedded in a generated file — or, when `name` is an `asset:` reference, what an *input* asset holds (`source` says which; `job` is null for an asset). Reading an input's duration, frame count, fps and sample rate before a run is how a caller learns the `total_frames`, `fps` and `sample_rate` a workflow expects it to supply: the exact workflow and arguments that produced it, and, for audio/video, a `media` block (duration, rate, channels, fps, size, peak/mean dBFS). `envelope=true` adds `media.envelope` — `rms_dbfs` and `peak_dbfs` one entry per second — which is what locates something in a track rather than measuring the whole of it. `workspace` names the workspace for this one call without switching the session to it - the same pin `run_workflow` takes, so a job run into another workspace stays reachable from the session that queued it |
 
 ### Media
@@ -249,12 +249,12 @@ The session starts in `default` and stays there unless it is told otherwise.
 
 | Tool | Arguments | Purpose |
 | --- | --- | --- |
-| `validate_workflow(workflow=None, name=None, workspace=None, arguments=None)` | exactly one of `workflow` (inline definition) or `name` (a stored workflow, as `list_workflows` reports it), optional `workspace`, optional `arguments` | Check a workflow against the schema and against real pipeline signatures. Free and instant. Validating by name uses the workflow file's own directory as the base directory, so it sees what a run would. Returns every schema violation in `errors`, each with the JSON path it sits at, so a draft is fixed in one pass, and a `previous_result:` that names no earlier step is one of them. `warnings` covers what still runs but is probably wrong - a signature mismatch, and, for a list-driven variable, an entry key no step reads, at the entry's path. `workspace` names the workspace for this one call without switching the session to it - use it to pin a job whose `output:` or `asset:` references live in a workspace other than the session's. Pass the same `arguments` you will pass to `run_workflow` and they are checked too - an undeclared or renamed variable name, a value that will not coerce to the declared type, and an `asset:`, `prompt:` or `output:` reference that names nothing this workspace can reach, each reported at `arguments.<name>`. `checked_arguments` lists what was covered, so a `valid: true` about the stored defaults cannot be mistaken for one about your values. A reference set the model would refuse - too many images, videos or audio clips, or, for MiniMax-H3, audio as the only reference - is an error here too, rather than a failure minutes into a run you acknowledged. `run_workflow` makes the same check and refuses a bad argument rather than queuing a job that fails on its first step. A valid answer carries `plan` - the fingerprint, step count, list lengths, `downloads_required` and `estimate` (with `basis`) for the arguments given; quote from it |
+| `validate_workflow(workflow=None, name=None, workspace=None, arguments=None)` | exactly one of `workflow` (inline definition) or `name` (a stored workflow, as `list_workflows` reports it), optional `workspace`, optional `arguments` | Check a workflow against the schema and against real pipeline signatures. Free and instant. Validating by name uses the workflow file's own directory as the base directory, so it sees what a run would. Returns every schema violation in `errors`, each with the JSON path it sits at, so a draft is fixed in one pass, and a `previous_result:` that names no earlier step is one of them. `warnings` covers what still runs but is probably wrong - a signature mismatch, and, for a list-driven variable, an entry key no step reads, at the entry's path. `workspace` names the workspace for this one call without switching the session to it - use it to pin a job whose `output:` or `asset:` references live in a workspace other than the session's. Pass the same `arguments` you will pass to `run_workflow` and they are checked too - an undeclared or renamed variable name, a value that will not coerce to the declared type, and an `asset:`, `prompt:` or `output:` reference that names nothing this workspace can reach, each reported at `arguments.<name>`. `checked_arguments` lists what was covered, so a `valid: true` about the stored defaults cannot be mistaken for one about your values. A reference set the model would refuse - too many images, videos or audio clips, or, for MiniMax-H3, audio as the only reference - is an error here too, rather than a failure minutes into a run you acknowledged. `run_workflow` makes the same check and refuses a bad argument rather than queuing a job that fails on its first step. A valid answer carries `plan` - the fingerprint, step count, list lengths, `downloads_required`, `estimate` (with `basis`) and `elided_steps` for the arguments given; quote from it. `steps` counts what will run: a step nothing reads and which saves no file does not run, and is named in `elided_steps` instead |
 | `list_workspaces()` | — | The server's workspaces and which one this session is using. Each has its own workflows, assets and outputs; the prompt library is shared by all of them |
 | `use_workspace(name)` | `name` | Work in that workspace for the rest of the session - every later call reads and writes there. This is how to keep your work out of another agent's namespace rather than sharing the default one. Checked against the server, so a typo fails here rather than scoping every later call to nothing |
 | `create_workspace(name, use=False)` | `name`, `use` | Create a workspace. Pass use=true to switch this session to it as well; otherwise the session stays where it was and the result says so |
 | `delete_workspace(name, acknowledged_cost=False)` | `name`, `acknowledged_cost` | Permanently delete a workspace and everything in it. Refuses without the acknowledgement, reporting what it would remove |
-| `list_assets()` | — | The input media on the server, each with the `asset:` reference a workflow argument carries. Look here before asking for a file - what a workflow needs may already be there |
+| `list_assets()` | — | The input media on the server, each with the `asset:` reference a workflow argument carries. Look here before asking for a file - what a workflow needs may already be there. `libraries` names the roots searched and which are writable; `shadowed` lists names a nearer library hides |
 | `keep_output(name, asset_name=None, overwrite=False, shared=False, workspace=None)` | `name`, optional `asset_name`, `overwrite`, `shared`, `workspace` | Keep a generated file as an input asset under a stable `asset:` name, so a later workflow can rely on it. The copy happens on the server: nothing is downloaded or re-uploaded. `asset_name` may name a folder and takes the kept file's extension when it has none; `shared=true` keeps it in the library every workspace shares, which is where a recurring cast belongs. `workspace` names the workspace for this one call without switching the session to it - the same pin `run_workflow` takes, so a job run into another workspace stays reachable from the session that queued it |
 | `upload_asset(file_path, asset_name=None, shared=False)` | `file_path` | Push a local image, video or audio file into the server's asset library and get back its `asset:` reference. The file is read from the machine the MCP server runs on, so this is how an input reaches a dw.serve running somewhere else. `asset_name` stores it under a readable name (`cast/priya-voice.wav`) instead of a random one; `shared=true` puts it in the library every workspace shares |
 | `delete_asset(name)` | `name` | Permanently remove one file from the asset library, by the name `list_assets` reports. Deletes from whichever library holds it - this workspace's own before the shared one; one from a read-only examples library is refused. Any workflow still carrying that `asset:` reference stops loading |