Available graphs

A list of the graphs available in agx. Files can be found in src/agx/_internal/known_graphs.

New naming convention rxx:

For the new graph_set: rxx, graph_type names are defined by separating the building blocks of different number of FGs by underscores and using hyphens to distinguish the multiplier by the number of FGs.

For example: rxx_4-4FG_6-2FG_4-1FG has 4 4FG building blocks, 6 2FG building blocks and 4 1FG building blocks.

Note that making new graphs is currently quite time consuming, so start with a smaller max_samples than the default for rxx, which is 1e6.

Graphs:

Important

All new graphs are run with a max_samples of 1e6.

Filtering graphs

The code no longer hard codes filtering within generation, but agx.TopologyCode offers methods for filtering for double wells or parallel edges. Additionally, one might want to build graphs that are not only one connected graph. To do so, you must change allowed_num_components.

Important

While we have one, two and three type graphs below, within each, any stoichiometry or configuration of same-numbered connections can be used.

One-type graphs

Produced graphs for m in (1 - 13) with FGs in (1 - 5). Generated with code:

bbs = {1, 2, 3, 4, 5}
multipliers = range(1, 11)
for midx, fgnum in it.product(multipliers, bbs):
    string = f"{midx}-{fgnum}FG"
    iterator = agx.TopologyIterator(
        node_counts={
            agx.NodeType(type_id=0, num_connections=fgnum): midx,
        },
        graph_set="rxx",
    )
    if iterator.is_type_possible():
        logger.info("trying %s", string)
        logger.info(
            "graph iteration has %s graphs", iterator.count_graphs()
        )

Not all graphs have been built due to cost, or non-valid chemistry.

# Missing.
[
    '1-1FG', '1-3FG', '1-5FG',
    '3-1FG', '3-3FG', '3-5FG',
    '5-1FG', '5-3FG', '5-5FG',
    '7-1FG', '7-3FG', '7-5FG',
    '9-1FG', '9-3FG', '9-5FG',
    '11-1FG', '11-3FG', '11-4FG', '11-5FG',
    '12-2FG', '12-3FG', '12-4FG', '12-5FG'
]

Two-type graphs

Produced graphs for m in (1 - 12) with FGs in (1 - 5) and a combinatorial set of stoichiometries of bigger:smaller. We limit the multiplier for stoichiometries other than 1:2. Generated with code:

bbs = {1, 2, 3, 4, 5}

# Two typers.
multipliers = range(1, 13)
two_type_stoichiometries = tuple(
    (i, j) for i, j in it.product((1, 2, 3, 4, 5), repeat=2)
)
for midx, fgnum1, fgnum2, stoich in it.product(
    multipliers, bbs, bbs, two_type_stoichiometries
):
    if fgnum1 == fgnum2:
        continue

    # Do not do all for larger stoichiomers.
    if stoich not in ((1, 2), ) and midx > 4:  # noqa: PLR2004
        continue

    fgnum1_, fgnum2_ = sorted((fgnum1, fgnum2), reverse=True)

    string = (
        f"{midx * stoich[0]}-{fgnum1_}FG_"
        f"{midx * stoich[1]}-{fgnum2_}FG"
    )

    iterator = agx.TopologyIterator(
        node_counts={
            agx.NodeType(type_id=0, num_connections=fgnum1_): midx
            * stoich[0],
            agx.NodeType(type_id=1, num_connections=fgnum2_): midx
            * stoich[1],
        },
        graph_set="rxx",
    )
    if iterator.is_type_possible():
        logger.info("trying %s", string)
        logger.info(
            "graph iteration has %s graphs", iterator.count_graphs()
        )

Not all graphs have been built due to cost, or non-valid chemistry.

# Missing.
[
    '8-5FG_20-2FG',
    '9-5FG_15-3FG',
    '12-4FG_16-3FG', '12-5FG_15-4FG','12-5FG_20-3FG', '12-4FG_24-2FG',
    '16-5FG_20-4FG',
]

Three-type graphs

Produced graphs for m in (1, 2) with FGs in (1 - 4) and a combinatorial check of stoichiometries. Note that current versions will always focus on smaller FG BBs binding only to the BB with the most FGs. Generated with code:

bbs = {1, 2, 3, 4, 5}

# Three typers.
multipliers = (1, 2)
three_type_stoichiometries = tuple(
    (i, j, k) for i, j, k in it.product((1, 2, 3, 4, 5, 6, 7, 8, 9, 10), repeat=3)
)
for midx, fgnum1, fgnum2, fgnum3, stoich in it.product(
    multipliers, bbs, bbs, bbs, three_type_stoichiometries
):
    if fgnum1 in (fgnum2, fgnum3) or fgnum2 == fgnum3:
        continue
    fgnum1_, fgnum2_, fgnum3_ = sorted(
        (fgnum1, fgnum2, fgnum3), reverse=True
    )

    string = (
        f"{midx * stoich[0]}-{fgnum1_}FG_"
        f"{midx * stoich[1]}-{fgnum2_}FG_"
        f"{midx * stoich[2]}-{fgnum3_}FG"
    )

    iterator = agx.TopologyIterator(
        node_counts={
            agx.NodeType(type_id=0, num_connections=fgnum1_): midx
            * stoich[0],
            agx.NodeType(type_id=1, num_connections=fgnum2_): midx
            * stoich[1],
            agx.NodeType(type_id=2, num_connections=fgnum3_): midx
            * stoich[2],
        },
        graph_set="rxx",
    )
    logger.info(
        "graph iteration has %s graphs", iterator.count_graphs()
    )
    if iterator.is_type_possible():
        logger.info("trying %s", string)
        logger.info(
            "graph iteration has %s graphs", iterator.count_graphs()
        )

Not all graphs have been built due to cost, or non-valid chemistry.

# Missing.
[
    "3-5FG_3-3FG_3-2FG",
    "4-5FG_4-4FG_2-2FG", "4-5FG_4-3FG_8-1FG", "4-5FG_8-2FG_4-1FG",
    "4-5FG_2-4FG_4-3FG", "4-5FG_6-3FG_2-1FG", "4-5FG_2-4FG_6-2FG",
    "4-5FG_3-4FG_4-2FG", "4-5FG_6-2FG_8-1FG", "4-4FG_2-3FG_5-2FG",
    "4-5FG_4-3FG_4-2FG",
    "5-5FG_4-4FG_3-3FG", "5-4FG_4-3FG_4-2FG", "5-5FG_5-3FG_5-2FG",
    "6-3FG_4-2FG_10-1FG", "6-4FG_10-2FG_4-1FG", "6-4FG_4-3FG_6-2FG",
    "6-5FG_6-4FG_6-1FG", "6-5FG_10-2FG_10-1FG", "6-5FG_6-3FG_6-2FG",
    "6-5FG_8-3FG_6-1FG", "6-5FG_6-4FG_2-3FG",
    "8-4FG_8-3FG_8-1FG", "8-4FG_10-3FG_2-1FG", "8-5FG_8-4FG_8-1FG",
    "8-3FG_10-2FG_4-1FG", "8-5FG_8-3FG_8-2FG", "8-4FG_8-3FG_4-2FG",
    "8-5FG_6-4FG_8-2FG", "8-5FG_10-3FG_10-1FG", "8-5FG_8-4FG_4-2FG",
    "8-5FG_4-4FG_8-3FG", "8-4FG_4-3FG_10-2FG",
    "10-5FG_10-4FG_10-1FG", "10-5FG_8-4FG_6-3FG", "10-4FG_10-3FG_10-1FG",
    "10-4FG_8-3FG_8-2FG", "10-5FG_10-3FG_10-2FG", "10-3FG_10-2FG_10-1FG",
    # And many more...
]