add funtions to cal angle to prevent sideroads getting values of main road

[wangzhao0217]

• update funtions get_vector and calculate_angle
• corr_line_geometry is not exist

[wangzhao0217]

Hi @Robinlovelace
Currently, I couldn't find a more efficient method to prevent sideroads from getting values from the main road, except to use the angle calculation method. I did the following update on file rnet_join.R

• add line in function rnet_join 109 rnet_y_centroids$corr_line_geometry = rnet_y$geometry
• add line in function rnet_merge 256 res_sf$angle = sapply(1:nrow(res_sf), function(i) {
  calculate_angle(get_vector(res_sf$corr_line_geometry[i,]), get_vector(res_sf$geometry[i,]))
  })
• updating

    for(i in sum_cols) {
      # TODO: deduplicate
      length_y = as.numeric(sf::st_length(rnet_y))
      mask <- (res_sf$angle < 15) | (res_sf$angle > 160)
      # i = sum_cols[1]
      res_sf[mask, i] <- res_sf[mask, i] / res_sf[mask, "length_x"]
      over_estimate <- sum(res_sf[mask, i] * res_sf[mask, "length_x"], na.rm = TRUE) /
        sum(rnet_y[mask, i] * length_y, na.rm = TRUE)
      
      res_sf[mask, i] <- res_sf[mask, i] / over_estimate
    }
get_vector <- function(line) {
  if (class(line) == "LINESTRING") {
    coords <- st_coordinates(line)
    start <- coords[1, 1:2]
    end <- coords[2, 1:2]
  } else {  # For MultiLineStrings, just use the first line
    first_line <- st_cast(line, "LINESTRING")[[1]]
    coords <- st_coordinates(first_line)
    start <- coords[1, 1:2]
    end <- coords[2, 1:2]
  }
  return(c(end[1] - start[1], end[2] - start[2]))
}

calculate_angle <- function(vector1, vector2) {
  dot_product <- sum(vector1 * vector2)
  magnitude_product <- sqrt(sum(vector1^2)) * sqrt(sum(vector2^2))
  cos_angle <- dot_product / magnitude_product
  angle <- acos(cos_angle) * (180 / pi)
  return(angle)
}

However, it is still not working, will keep fixing it.

[Robinlovelace]

Thanks Zhao, will take a look.

[Robinlovelace]

Looks good Zhao. Do you have any results showing the removal of high values on side roads?

[Robinlovelace]

These are hard coded. At some point it would be good to make these available as arguments.

[wangzhao0217]

These are hard coded. At some point it would be good to make these available as arguments.

Got it, Robin. Sorry at the moment there is a bug in the code, I am working on it to solve it. (Switching from Python to R for programming is tough that I thought)

[Robinlovelace]

Yes it's tricky. Keep up the good work and no worries about making these hardcoded values available as arguments, can add later..

[wangzhao0217]

Hi @Robinlovelace got some results, catch up with you tomorrow

[Robinlovelace]

Looking good. Let's discuss and hopefully merge this today @wangzhao0217

[wangzhao0217]

will make new pr

[Robinlovelace]

Sounds good, looking forward to reviewing it and seeing the results.