Fix NA cluster ID values

[russHyde]

Using "scanner-env-2022-08-01.rds" from https://github.com/oliviabboyd/tfpscan_sample .

tree <- treeview(
  "tfpscan-2022-08-01/scanner-env-2022-08-01.rds",
  output_dir = "tfpscan-2022-08-01/treeview"
)

When running the above in the trim-tooltip branch, most of the tooltip text are of the form "Cluster ID: 12345", but some are of the form "Cluster ID: NA". This corresponds to tree$data$mouseover. But the relevant rows have tree$data$cluster_id matching NA_character_.

[russHyde]

In the following code from treeview()

  sc2 <- sc0[!is.na(sc0$tr2mrca), ]
  sc2$date_range <- sapply(
    seq_len(nrow(sc2)),
    function(i) glue::glue("{sc2$least_recent_tip[i]} -> {sc2$most_recent_tip[i]}")
  )

  ## tips
  td0 <- sc2[sc2$tr2mrca <= ape::Ntip(tr2), tdvars]
  td0$lineages <- td0$lineage
  td0$cocirc_summary <- td0$cocirc_lineage_summary
  td0$node <- td0$tr2mrca
  td0$internal <- "N"

  ## internal
  td1 <- sc2[sc2$tr2mrca > ape::Ntip(tr2), tdvars]
  if (nrow(td1) > 0) {
    td1$lineages <- td1$lineage
    td1$cocirc_summary <- td1$cocirc_lineage_summary
    td1$node <- td1$tr2mrca
    td1$internal <- "Y"
    td1$cluster_size <- 0
    x <- setdiff(
      (ape::Ntip(tr2) + 1):(ape::Ntip(tr2) + ape::Nnode(tr2)),
      td1$node
    ) # make sure every node represented
    td1 <- merge(td1,
      data.frame(node = x),
      all = TRUE
    )
    td <- rbind(td0, td1)
  } else {
    td <- td0
  }
  td <- td[order(td$node), ] # important

The sc0 data.frame from which tree$data is ultimately formed, has no NA cluster_id entries.
Nor are there NA cluster_ids in sc2, td0 or td1 (prior to definition of x).

[russHyde]

But, we have

x == c(175, 181, 188, 189, 190, 199)

And after the merge(td1, data.frame(node = x), all = TRUE) we do have NA values in cluster_id

td1$cluster_id
 [1] NA       "94121"  "94137"  "94217"  "96096"  "96097"  NA       "97961"  "97988"  "98462"  "98463"  "98464" 
[13] "99556"  NA       NA       NA       "100556" "100557" "100558" "100564" "101069" "102175" "102216" "102424"
[25] NA       "102637" "103258" "103913" "104820" "105030" "106973" "106999" "107001" "108938" "108939" "108940"
[37] "108947" "109088" "109306" "109307" "109308" "110040" "110519" "111323" "111355" "113060" "113508" "114073"

[russHyde]

In the above,

• td0$node contains the sequence 1 .. 174.
• prior to merging with x, td1$node contains

sort(td1$node)
 [1] 176 177 178 179 180 182 183 184 185 186 187 191 192 193 194 195 196 197 198 200 201 202 203 204 205 206 207 208
[29] 209 210 211 212 213 214 215 216 217 218 219 220 221 222

[russHyde]

The tips/leaves of the tree are the nodes with ID 1 .. ape::Ntip(tr2), where here ape::Ntip(tr2) == 174

The internal nodes appear to be the nodes with ID (ntip + 1) : (ntip + nnode), where ntip = ape::Ntip(tr2) and nnode = ape::Nnode(tr2).

[russHyde]

node values derive from sc0$tr2mrca.
The code to add tr2mrca to sc0 for internal nodes looks like this:

  for (i in seq_along(stres2)) {
    inode <- i + ape::Ntip(tr2)
    uv <- stats::na.omit(sc0$node_number[match(
      stres2[[i]]$tip.label,
      sc0$representative
    )])
    shared_anc <- Reduce(
      intersect,
      e0$ancestors[uv]
    )
    shared_anc2 <- setdiff(
      intersect(
        shared_anc,
        sc0$node_number
      ),
      uv
    )
    if (length(shared_anc2) > 0) {
      a <- shared_anc2[which.min(e0$ndesc[shared_anc2])]
      sc0$tr2mrca[sc0$node_number == a] <- inode
    }
  }

[russHyde]

for inode = 175, the shared_anc2 vector is empty. OMG there must be a simple way to set up a mapping between tree node and cluster ID.