DataLog

class wpiutil.log.DataLog(*args, **kwargs)

Bases: pybind11_object

A data log. The log file is created immediately upon construction with a temporary filename. The file may be renamed at any time using the SetFilename() function.

The lifetime of the data log object must be longer than any data log entry objects that refer to it.

The data log is periodically flushed to disk. It can also be explicitly flushed to disk by using the Flush() function.

Finish() is needed only to indicate in the log that a particular entry is no longer being used (it releases the name to ID mapping). Finish() is not required to be called for data to be flushed to disk; entries in the log are written as Append() calls are being made. In fact, Finish() does not need to be called at all; this is helpful to avoid shutdown races where the DataLog object might be destroyed before other objects. It’s often not a good idea to call Finish() from destructors for this reason.

DataLog calls are thread safe. DataLog uses a typical multiple-supplier, single-consumer setup. Writes to the log are atomic, but there is no guaranteed order in the log when multiple threads are writing to it; whichever thread grabs the write mutex first will get written first. For this reason (as well as the fact that timestamps can be set to arbitrary values), records in the log are not guaranteed to be sorted by timestamp.

Overloaded function.

  1. __init__(self: wpiutil._wpiutil.log.DataLog, dir: str = ‘’, filename: str = ‘’, period: float = 0.25, extraHeader: str = ‘’) -> None

Construct a new Data Log. The log will be initially created with a temporary filename.

Parameters:

  • dir – directory to store the log

  • filename – filename to use; if none provided, a random filename is generated of the form “wpilog_{}.wpilog”

  • period – time between automatic flushes to disk, in seconds; this is a time/storage tradeoff

  • extraHeader – extra header data

  1. __init__(self: wpiutil._wpiutil.log.DataLog, write: Callable[[Buffer], None], period: float = 0.25, extraHeader: str = ‘’) -> None

Construct a new Data Log that passes its output to the provided function rather than a file. The write function will be called on a separate background thread and may block. The write function is called with an empty data array when the thread is terminating.

Parameters:

  • write – write function

  • period – time between automatic calls to write, in seconds; this is a time/storage tradeoff

  • extraHeader – extra header data

addSchema(*args, **kwargs)

Overloaded function.

  1. addSchema(self: wpiutil._wpiutil.log.DataLog, name: str, type: str, schema: Buffer, timestamp: int = 0) -> None

Registers a data schema. Data schemas provide information for how a certain data type string can be decoded. The type string of a data schema indicates the type of the schema itself (e.g. “protobuf” for protobuf schemas, “struct” for struct schemas, etc). In the data log, schemas are saved just like normal records, with the name being generated from the provided name: “/.schema/<name>”. Duplicate calls to this function with the same name are silently ignored.

Parameters:

  • name – Name (the string passed as the data type for records using this schema)

  • type – Type of schema (e.g. “protobuf”, “struct”, etc)

  • schema – Schema data

  • timestamp – Time stamp (may be 0 to indicate now)

  1. addSchema(self: wpiutil._wpiutil.log.DataLog, name: str, type: str, schema: str, timestamp: int = 0) -> None

Registers a data schema. Data schemas provide information for how a certain data type string can be decoded. The type string of a data schema indicates the type of the schema itself (e.g. “protobuf” for protobuf schemas, “struct” for struct schemas, etc). In the data log, schemas are saved just like normal records, with the name being generated from the provided name: “/.schema/<name>”. Duplicate calls to this function with the same name are silently ignored.

Parameters:

  • name – Name (the string passed as the data type for records using this schema)

  • type – Type of schema (e.g. “protobuf”, “struct”, etc)

  • schema – Schema data

  • timestamp – Time stamp (may be 0 to indicate now)

addStructSchema(type: type, timestamp: int = 0) → None

Registers a struct schema. Duplicate calls to this function with the same name are silently ignored.

@tparam T struct serializable type

Parameters:

  • type – optional struct type info

  • timestamp – Time stamp (0 to indicate now)

appendBoolean(entry: int, value: bool, timestamp: int) → None

Appends a boolean record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • value – Boolean value to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendBooleanArray(entry: int, arr: List[bool], timestamp: int) → None

Appends a boolean array record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • arr – Boolean array to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendDouble(entry: int, value: float, timestamp: int) → None

Appends a double record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • value – Double value to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendDoubleArray(entry: int, arr: List[float], timestamp: int) → None

Appends a double array record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • arr – Double array to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendFloat(entry: int, value: float, timestamp: int) → None

Appends a float record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • value – Float value to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendFloatArray(entry: int, arr: List[float], timestamp: int) → None

Appends a float array record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • arr – Float array to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendInteger(entry: int, value: int, timestamp: int) → None

Appends an integer record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • value – Integer value to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendIntegerArray(entry: int, arr: List[int], timestamp: int) → None

Appends an integer array record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • arr – Integer array to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendRaw(entry: int, data: Buffer, timestamp: int) → None

Appends a raw record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • data – Byte array to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendRaw2(entry: int, data: List[Buffer], timestamp: int) → None

Appends a raw record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • data – Byte array to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendString(entry: int, value: str, timestamp: int) → None

Appends a string record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • value – String value to record

  • timestamp – Time stamp (may be 0 to indicate now)

appendStringArray(entry: int, arr: List[str], timestamp: int) → None

Appends a string array record to the log.

Parameters:

  • entry – Entry index, as returned by Start()

  • arr – String array to record

  • timestamp – Time stamp (may be 0 to indicate now)

finish(entry: int, timestamp: int = 0) → None

Finish an entry.

Parameters:

  • entry – Entry index

  • timestamp – Time stamp (may be 0 to indicate now)

flush() → None

Explicitly flushes the log data to disk.

hasSchema(name: str) → bool

Returns whether there is a data schema already registered with the given name.

Parameters:

name – Name (the string passed as the data type for records using this schema)

Returns:

True if schema already registered

pause() → None

Pauses appending of data records to the log. While paused, no data records are saved (e.g. AppendX is a no-op). Has no effect on entry starts / finishes / metadata changes.

resume() → None

Resumes appending of data records to the log. If called after Stop(), opens a new file (with random name if SetFilename was not called after Stop()) and appends Start records and schema data values for all previously started entries and schemas.

setFilename(filename: str) → None

Change log filename.

Parameters:

filename – filename

setMetadata(entry: int, metadata: str, timestamp: int = 0) → None

Updates the metadata for an entry.

Parameters:

  • entry – Entry index

  • metadata – New metadata for the entry

  • timestamp – Time stamp (may be 0 to indicate now)

start(name: str, type: str, metadata: str = '', timestamp: int = 0) → int

Start an entry. Duplicate names are allowed (with the same type), and result in the same index being returned (Start/Finish are reference counted). A duplicate name with a different type will result in an error message being printed to the console and 0 being returned (which will be ignored by the Append functions).

Parameters:

  • name – Name

  • type – Data type

  • metadata – Initial metadata (e.g. data properties)

  • timestamp – Time stamp (may be 0 to indicate now)

Returns:

Entry index

stop() → None

Stops appending all records to the log, and closes the log file.