ray.rllib.env.single_agent_episode.SingleAgentEpisode.get_infos

SingleAgentEpisode.get_infos(indices: int | slice | List[int] | None = None, *, neg_index_as_lookback: bool = False, fill: Any | None = None) → Any

Returns individual info dicts or list (ranges) thereof from this episode.

Parameters:

• indices – A single int is interpreted as an index, from which to return the individual info dict stored at this index. A list of ints is interpreted as a list of indices from which to gather individual info dicts in a list of size len(indices). A slice object is interpreted as a range of info dicts to be returned. Thereby, negative indices by default are interpreted as “before the end” unless the neg_index_as_lookback=True option is used, in which case negative indices are interpreted as “before ts=0”, meaning going back into the lookback buffer. If None, will return all infos (from ts=0 to the end).

• neg_index_as_lookback – If True, negative values in indices are interpreted as “before ts=0”, meaning going back into the lookback buffer. For example, an episode with infos [{“l”:4}, {“l”:5}, {“l”:6}, {“a”:7}, {“b”:8}, {“c”:9}], where the first 3 items are the lookback buffer (ts=0 item is {“a”: 7}), will respond to get_infos(-1, neg_index_as_lookback=True) with {"l":6} and to get_infos(slice(-2, 1), neg_index_as_lookback=True) with [{"l":5}, {"l":6},  {"a":7}].

• fill – An optional value to use for filling up the returned results at the boundaries. This filling only happens if the requested index range’s start/stop boundaries exceed the episode’s boundaries (including the lookback buffer on the left side). This comes in very handy, if users don’t want to worry about reaching such boundaries and want to auto-fill. For example, an episode with infos [{“l”:10}, {“l”:11}, {“a”:12}, {“b”:13}, {“c”:14}] and lookback buffer size of 2 (meaning infos {“l”:10}, {“l”:11} are part of the lookback buffer) will respond to get_infos(slice(-7, -2), fill={"o": 0.0}) with [{"o":0.0}, {"o":0.0}, {"l":10}, {"l":11}, {"a":12}].

Examples:

from ray.rllib.env.single_agent_episode import SingleAgentEpisode

episode = SingleAgentEpisode(
    infos=[{"a":0}, {"b":1}, {"c":2}, {"d":3}],
    # The following is needed, but not relevant for this demo.
    observations=[0, 1, 2, 3], actions=[1, 2, 3], rewards=[1, 2, 3],
    len_lookback_buffer=0,  # no lookback; all data is actually "in" episode
)
# Plain usage (`indices` arg only).
episode.get_infos(-1)  # {"d":3}
episode.get_infos(0)  # {"a":0}
episode.get_infos([0, 2])  # [{"a":0},{"c":2}]
episode.get_infos([-1, 0])  # [{"d":3},{"a":0}]
episode.get_infos(slice(None, 2))  # [{"a":0},{"b":1}]
episode.get_infos(slice(-2, None))  # [{"c":2},{"d":3}]
# Using `fill=...` (requesting slices beyond the boundaries).
# TODO (sven): This would require a space being provided. Maybe we can
#  skip this check for infos, which don't have a space anyways.
# episode.get_infos(slice(-5, -3), fill={"o":-1})  # [{"o":-1},{"a":0}]
# episode.get_infos(slice(3, 5), fill={"o":-2})  # [{"d":3},{"o":-2}]

Returns:

The collected info dicts. As a 0-axis batch, if there are several indices or a list of exactly one index provided OR indices is a slice object. As single item (B=0 -> no additional 0-axis) if indices is a single int.