Whereas studying about fee channels, I understood channel attributes like base charges, proportional charges and most HTLC dimension could be directed, i.e., the attributes from A to B wouldn’t essentially be the identical because the attributes from B to A. Nevertheless, as I used to be taking part in round with this Lightning dataset by Christian Decker, I seen most channel information do not need a corresponding entry for the attributes in the wrong way:

  <key id="d17" for="edge" attr.identify="cltv_expiry_delta" attr.sort="int" />
  <key id="d16" for="edge" attr.identify="htlc_maximum_msat" attr.sort="int" />
  <key id="d15" for="edge" attr.identify="htlc_minimim_msat" attr.sort="int" />
  <key id="d14" for="edge" attr.identify="fee_proportional_millionths" attr.sort="int" />
  <key id="d13" for="edge" attr.identify="fee_base_msat" attr.sort="int" />
  <key id="d12" for="edge" attr.identify="options" attr.sort="string" />
  <key id="d11" for="edge" attr.identify="timestamp" attr.sort="int" />
  <key id="d10" for="edge" attr.identify="vacation spot" attr.sort="string" />
  <key id="d9" for="edge" attr.identify="supply" attr.sort="string" />
  <key id="d8" for="edge" attr.identify="scid" attr.sort="string" />

<edge supply="020bbb584a5008a9d1dfcf9aca44b254919d9f33e37e8dbdba9e0dab3cbfc628dc" goal="029a8741675c4c9078b577ddc4348d602d2fb45a12e6087b617925997a84f4c02e">
  <information key="d8">531987x85x0/0</information>
  <information key="d9">020bbb584a5008a9d1dfcf9aca44b254919d9f33e37e8dbdba9e0dab3cbfc628dc</information>
  <information key="d10">029a8741675c4c9078b577ddc4348d602d2fb45a12e6087b617925997a84f4c02e</information>
  <information key="d11">1578286721</information>
  <information key="d12" />
  <information key="d13">1000</information>
  <information key="d14">1</information>
  <information key="d15">1000</information>
  <information key="d16">87340000</information>
  <information key="d17">144</information>
</edge>

// No edge from 029a8741675c4c9078b577ddc4348d602d2fb45a12e6087b617925997a84f4c02e to 020bbb584a5008a9d1dfcf9aca44b254919d9f33e37e8dbdba9e0dab3cbfc628dc within the dataset

If the attributes are directed, I anticipated to have an edge (j,i) for each edge (i,j) within the dataset. Is that this an issue of the dataset? May the attributes be legitimate for each instructions?