fluxy-ign 0.1.1

Python client for controlled Ignition Gateway automation through WebDev

fluxy

A Python library that emulates useful subsets of Ignition's system APIs from outside Ignition.

Install as a library:

uv add fluxy-ign

Install as a deployment tool:

uv tool install fluxy-ign

Install with MCP support:

uv tool install 'fluxy-ign[mcp]'

Authenticated WebDev deployment:

fluxy-deploy-webdev /path/to/ignition/data/projects/flux \
  --auth-token-file /path/to/fluxy-token

Clients must pass the same bearer token:

fx = Fluxy("http://localhost:8088/system/webdev/flux", token="shared-secret-token")

See docs/auth.md for the full auth workflow, including token creation, project scan, client configuration, and troubleshooting 401/403 responses.

Run the optional MCP adapter:

fluxy-mcp \
  --base-url http://localhost:8088/system/webdev/flux \
  --token-file /path/to/fluxy-token

The MCP server is read-only by default. Add --allow-writes to expose write/configure tools and --allow-destructive to expose destructive tools such as tag deletion.

The PyPI distribution is fluxy-ign; the Python import remains fluxy.

Start with docs/user-guide.md for the user workflow. See docs/ for architecture, deployment, API, Gateway config, and integration-test details.

Initial scope:

• fx.tag.read_blocking(tag_paths)
• fx.tag.write_blocking(tag_paths, values)
• fx.tag.copy(tag_paths, destination_path)
• fx.tag.configure(tags, base_path=None, collision_policy="o")
• fx.tag.browse(path=None, tag_filter=None)
• fx.tag.query(provider, query=None, limit=None, continuation=None)
• fx.alarm helpers for tested disposable alarm query/shelve/unshelve/acknowledge workflows
• fx.db helpers for Ignition datasource management, SQL queries, prepared queries, updates, transactions, and named queries
• fx.device helpers for tested Ignition simulator-device add/list/enable/remove workflows
• fx.historian helpers for tested Core Historian browse and raw point injection/query workflows
• fx.opc helpers for tested disposable simulator-device OPC browse/read/write workflows
• fx.report helpers for tested report listing/execution workflows
• fx.user helpers for tested disposable user/role workflows against userDB
• fx.util helpers for tested gateway diagnostics and audit write/query workflows
• fx.project.request_scan()
• python -m fluxy.gateway_config for narrow Gateway SQLite/PostgreSQL connection resource deployment
• fluxy-mcp optional MCP server over the Fluxy API, read-only by default
• HTTP bridge to Ignition WebDev endpoints that call the real system.tag, tested system.alarm, system.db, tested system.device, tested system.historian, tested system.opc, and tested system.util APIs inside the gateway.

Snake_case is canonical for Python code. Ignition-style aliases are available for porting scripts:

• fx.tag.readBlocking(...)
• fx.tag.writeBlocking(...)
• fx.project.requestScan()

Python Usage

from fluxy import Fluxy

fx = Fluxy(
    base_url="https://ignition.example.com/system/webdev/<webdev-project>",
    project_location="/usr/local/bin/ignition/data/projects/flux",
    tag_provider="default",
    token="shared-secret-token",
)

values = fx.tag.read_blocking([
    "[default]Path/To/Tag1",
    "[default]Path/To/Tag2",
])

for value in values:
    print(value.tag_path, value.value, value.quality, value.timestamp)

Single-tag reads also work and return a single QualifiedValue:

value = fx.tag.read_blocking("[default]Path/To/Tag1")

Write/readback shape:

tag_paths = [
    "[default]Demo/Area_A/Unit_01/LOAD_FACTOR",
]

qualities = fx.tag.write_blocking(tag_paths, [1.1])
values = fx.tag.read_blocking(tag_paths)

Single-tag writes also work and return a single WriteResult:

quality = fx.tag.write_blocking("[default]Path/To/Setpoint", 12.3)

Configure shape:

qualities = fx.tag.configure(
    [
        {
            "name": "MemoryFloat",
            "tagType": "AtomicTag",
            "valueSource": "memory",
            "dataType": "Float4",
            "value": 1.0,
        }
    ],
    base_path="[default]Folder",
    collision_policy="o",
)

Browse shape:

results = fx.tag.browse("[default]Folder")
for result in results:
    print(result.name, result.full_path, result.tag_type, result.data_type)

Tag discovery shape:

missing = fx.tag.read_blocking("[default]Folder/MissingTag")
if str(missing.quality).startswith("Bad_DoesNotExist"):
    print("tag is missing")

fx.tag.copy("[default]Folder/MemoryFloat", "[default]Folder/Copies")

results = fx.tag.query("default", query={"condition": {"path": "*MemoryFloat*"}})

system.tag.exists is forbidden in Fluxy. Use read_blocking(...) and monitor for Bad_DoesNotExist instead; it is more performant because it stays on the normal tag-read path instead of adding a dedicated gateway call.

Alarm shape:

source = "prov:default:/tag:FluxyAlarmIntegration/AlarmFloat:/alm:HighAlarm"
rows = fx.alarm.query_status(source=[source], include_shelved=True)
fx.alarm.shelve([source], timeout_seconds=60)
shelved = fx.alarm.get_shelved_paths()
fx.alarm.unshelve([source])

Dataset-like database results return QueryResult, a list[dict[str, Any]] subclass with boundary metadata. Ignition Datasets are serialized as columns + rows by WebDev and converted to row mappings in Python. See docs/user-guide.md#dataset-results.

Live DB tests cover 256-row mixed SQLite datasets. Integers/reals/text/nulls preserve JSON-compatible values, SQLite boolean-like values return 0/1, epoch milliseconds remain integers, and BLOBs return byte-value lists.

Device management shape:

fx.device.add_device("Simulator", "FluxyTemporarySimulator", {"Enabled": 0})
devices = fx.device.list_devices()
fx.device.set_device_enabled("FluxyTemporarySimulator", False)
fx.device.remove_device("FluxyTemporarySimulator")

Historian shape:

path = "histprov:Core Historian:/sys:gateway:/prov:default:/tag:FluxyHistorianIntegration/Test"
fx.historian.store_data_points([path], [12.5], [1778545000000], [192])
rows = fx.historian.query_raw_points([path], 1778544990000, 1778545010000)
aggregated = fx.historian.query_aggregated_points(
    [path],
    1778544990000,
    1778545010000,
    aggregates=["Maximum"],
    column_names=["value"],
)
browse_results = fx.historian.browse(
    "histprov:Core Historian:/sys:gateway:/prov:default:/tag:FluxyHistorianIntegration"
)
fx.historian.store_metadata([path], [1778545000000], {"documentation": "checked"})
metadata = fx.historian.query_metadata([path], start_date=1778544990000, end_date=1778545010000)
qualities = fx.historian.store_annotations(
    [path], [1778544995000], end_times=[1778545005000], types=["note"], data=["checked"]
)
annotations = fx.historian.query_annotations([path], 1778544990000, end_date=1778545010000)
fx.historian.delete_annotations([path], [annotations[0].storage_id])

The historian WebDev bridge selects system.historian.* on Ignition 8.3+ and falls back to the 8.1 system.tag.* historian functions when needed. Prefer 8.3 sys/prov historical paths in Python; the 8.1 storage fallback maps them to legacy storeTagHistory arguments.

Live tag-to-history boundary tests show Core Historian raw data points are numeric-oriented: Boolean stores as 1.0/0.0, numeric tags round-trip numerically, DateTime stores as epoch milliseconds, String does not query back through queryRawPoints despite a Good store quality, and Document is rejected by storeDataPoints.

OPC shape:

server = "Ignition OPC UA Server"
servers = fx.opc.get_servers(include_disabled=True)
rows = fx.opc.browse(opc_server=server, device="FluxyOpcSimulator")
device_nodes = fx.opc.browse_server(server, "Devices")
simple_rows = fx.opc.browse_simple(opc_server=server, device="FluxyOpcSimulator")
value = fx.opc.read_value(server, rows[0]["opcItemPath"])
quality = fx.opc.write_value(server, rows[0]["opcItemPath"], 123)

Util diagnostics and audit shape:

version = fx.util.get_version()
modules = fx.util.get_modules()
fx.util.audit("FluxyIntegrationAudit", action_target="unique", audit_profile="Audit")
rows = fx.util.query_audit_log("Audit", action_filter="FluxyIntegrationAudit")

fx.util.get_version() is cached after the first gateway read. Use fx.util.refresh_version() after changing gateway versions.

Report shape:

project = fx.project.get_project_name()
reports = fx.report.get_report_names_as_list(project)
report_rows = fx.report.get_report_names_as_dataset(project)
pdf = fx.report.execute_report("test_Report", project, file_type="pdf")

User source shape:

sources = fx.user.get_user_sources()
fx.user.add_role("UserDB", "fluxy_role")
fx.user.edit_role("UserDB", "fluxy_role", "fluxy_role_edited")
fx.user.add_user("UserDB", "fluxy_user", "password", roles=["fluxy_role"])
user = fx.user.get_user("UserDB", "fluxy_user")
fx.user.remove_user("UserDB", "fluxy_user")
fx.user.remove_role("UserDB", "fluxy_role_edited")
fx.user.add_schedule("fluxy_schedule", source_schedule="Always")
fx.user.remove_schedule("fluxy_schedule")
fx.user.add_holiday("fluxy_holiday", 2114904400000)
fx.user.remove_holiday("fluxy_holiday")

Project scan shape:

result = fx.project.request_scan()
print(result.ok, result.message)

WebDev Contract

Read request:

{
  "tagPaths": ["[default]Path/To/Tag"],
  "timeoutMs": 45000
}

tagPaths is the canonical request key. The Ignition WebDev bridge also accepts tag_paths and tag_list as compatibility aliases for existing generated files.

Configure request:

{
  "basePath": "[default]Folder",
  "tags": [
    {
      "name": "MemoryFloat",
      "tagType": "AtomicTag",
      "valueSource": "memory",
      "dataType": "Float4",
      "value": 1.0
    }
  ],
  "collisionPolicy": "o"
}

Read response:

{
  "ok": true,
  "values": [
    {
      "tagPath": "[default]Path/To/Tag",
      "value": 1.23,
      "quality": "Good",
      "timestamp": "2026-05-11T12:00:00.000Z"
    }
  ]
}

Install For Local Development

uv sync
uv run pytest

Ignition Setup

Create a WebDev project named Fluxy or adjust base_url to your chosen path.

Deploy the package-owned WebDev resource library into an Ignition project with:

uv run python -m fluxy.deploy_webdev /path/to/ignition/data/projects/flux

The WebDev resources are grouped under fluxy/ so this library can be tracked independently of other WebDev resources:

• /fluxy/tag/readBlocking
• /fluxy/tag/writeBlocking
• /fluxy/tag/copy
• /fluxy/tag/configure
• /fluxy/tag/browse
• /fluxy/tag/queryTags
• /fluxy/alarm/queryStatus
• /fluxy/alarm/shelve
• /fluxy/alarm/unshelve
• /fluxy/alarm/getShelvedPaths
• /fluxy/alarm/acknowledge
• /fluxy/db/...
• /fluxy/device/listDevices
• /fluxy/device/addDevice
• /fluxy/device/removeDevice
• /fluxy/device/setDeviceEnabled
• /fluxy/historian/browse
• /fluxy/historian/storeDataPoints
• /fluxy/historian/queryRawPoints
• /fluxy/historian/queryAggregatedPoints
• /fluxy/historian/storeAnnotations
• /fluxy/historian/queryAnnotations
• /fluxy/historian/deleteAnnotations
• /fluxy/historian/storeMetadata
• /fluxy/historian/queryMetadata
• /fluxy/opc/getServers
• /fluxy/opc/getServerState
• /fluxy/opc/browse
• /fluxy/opc/browseServer
• /fluxy/opc/browseSimple
• /fluxy/opc/readValue
• /fluxy/opc/readValues
• /fluxy/opc/writeValue
• /fluxy/opc/writeValues
• /fluxy/util/getVersion
• /fluxy/util/getModules
• /fluxy/util/getGatewayStatus
• /fluxy/util/getProjectName
• /fluxy/util/audit
• /fluxy/util/queryAuditLog
• /fluxy/report/getReportNamesAsList
• /fluxy/report/getReportNamesAsDataset
• /fluxy/report/executeReport
• /fluxy/user/getUserSources
• /fluxy/user/getRoles
• /fluxy/user/addRole
• /fluxy/user/editRole
• /fluxy/user/removeRole
• /fluxy/user/addUser
• /fluxy/user/getUser
• /fluxy/user/getUsers
• /fluxy/user/editUser
• /fluxy/user/removeUser
• /fluxy/user/addSchedule
• /fluxy/user/getSchedule
• /fluxy/user/getSchedules
• /fluxy/user/removeSchedule
• /fluxy/user/addHoliday
• /fluxy/user/getHoliday
• /fluxy/user/getHolidays
• /fluxy/user/removeHoliday
• /fluxy/project/requestScan

If editing the WebDev doPost section manually in Designer, use the body-only script instead:

ignition_webdev/designer_body/readBlocking_doPost_body.py

Designer adds def doPost(request, session): automatically. Do not paste a second function definition into that editor.

If AUTH_TOKEN is set in the WebDev scripts, configure the same token in Python:

from fluxy import Fluxy

fx = Fluxy(base_url="https://host/system/webdev/<webdev-project>", token="same-token")

Security note: writeBlocking is powerful. The current integration tests create disposable memory tags. Before using this against real operational tags, add gateway-side auth plus an allowlist.

There is also a controlled memory-tag write/readback probe for local trials:

uv run python scripts/write_readback_generated_tags.py \
  --base-url "http://localhost:8088/system/webdev/flux"

Integration Trial

The generated tag probe reads a tiny sample of full Ignition tag paths through the live WebDev bridge.

Generated-tag interface contract:

• Input files provide full Ignition tag path strings.
• Paths include providers, for example [default]Demo/Area_A/.../LOAD_FACTOR.
• fluxy sends them to WebDev as JSON under the canonical tagPaths key.
• WebDev runs system.tag.readBlocking(tag_paths, timeout_ms) inside Ignition and returns value, quality, and timestamp per path.

1. Deploy the WebDev resources with fluxy.deploy_webdev, then call fx.project.request_scan().

2. Confirm the project exposes:

https://host/system/webdev/<webdev-project>/fluxy/tag/readBlocking
https://host/system/webdev/<webdev-project>/fluxy/tag/writeBlocking
https://host/system/webdev/<webdev-project>/fluxy/tag/configure
https://host/system/webdev/<webdev-project>/fluxy/tag/browse
https://host/system/webdev/<webdev-project>/fluxy/tag/queryTags
https://host/system/webdev/<webdev-project>/fluxy/db/getConnections
https://host/system/webdev/<webdev-project>/fluxy/project/requestScan

If /system/webdev/<webdev-project> returns Resource "null" not found, that only means the project root was reached without a resource name. Test /fluxy/tag/readBlocking instead.

3. Run the generated-tag probe:

uv run python scripts/read_generated_tags.py \
  --base-url "https://host/system/webdev/<webdev-project>" \
  --sample-size 3

Optional environment-based version:

FLUXY_BASE_URL="https://host/system/webdev/<webdev-project>" \
FLUXY_SAMPLE_SIZE=3 \
uv run python scripts/read_generated_tags.py

The normal pytest suite includes closed-loop live gateway integration tests:

FLUXY_BASE_URL="https://host/system/webdev/<webdev-project>" \
uv run pytest tests/test_integration_generated_tags.py

The write/readback test creates disposable memory tags and removes them during cleanup:

FLUXY_BASE_URL="https://host/system/webdev/<webdev-project>" \
uv run pytest tests/test_integration_generated_write_readback.py

The configure integration test creates memory tags, reads their configured values, writes new values, and reads them back:

FLUXY_BASE_URL="https://host/system/webdev/<webdev-project>" \
FLUXY_CONFIGURE_BASE_PATH="[default]FluxyDemo" \
uv run pytest tests/test_integration_configure_types.py

The browse integration test inspects those configured memory tags:

FLUXY_BASE_URL="https://host/system/webdev/<webdev-project>" \
FLUXY_CONFIGURE_BASE_PATH="[default]FluxyDemo" \
uv run pytest tests/test_integration_browse_configured_tags.py

Useful optional variables:

• FLUXY_TOKEN: bearer token if AUTH_TOKEN is configured in WebDev.
• FLUXY_TAG_PATHS_FILE: override the generated tag path file.
• FLUXY_SAMPLE_SIZE: number of tag paths to read, default 3.
• FLUXY_TIMEOUT_MS: Ignition read timeout, default 45000.