dt <- read_parquet(here::here("data-processed", 
                              "misto_fix-02-gnames.parquet"))
dtl <- read_parquet(here::here("data-processed", 
                               "misto_fix-02-gnames_long-geo.parquet"))

ops <- read_parquet(here::here("data-processed", "op-codes.parquet"))
prj_meta <- read_parquet(here::here("data-processed", "prj-esif-meta.parquet"))

dt_chkonp <- read_parquet(here::here("data-processed", "prj_chkonp.parquet"))

prj_no_obec <- read_parquet(here::here("data-processed", "prj_id_noobec.parquet"))
dt_geostructure_bylvl <- read_parquet(here::here("data-processed", "dt_geostructure_by-lvl.parquet"))
dt_geostructure_byprj <- read_parquet(here::here("data-processed", "dt_geostructure_by-prj.parquet"))

source(here::here("read_metadata.R"))

zuj_geo <- read_rds(here::here("data-processed", "geo_zuj-all.rds"))

chu_geo <- CzechData::load_Data50("ChraneneUzemi") %>%
  group_by(KODCHU, NAZEV, KATEGCHU) %>% 
  summarise(pocet = n()) %>% 
  st_cast("MULTIPOLYGON") %>% 
  st_as_sf() %>% 
  ungroup()

# mapview::mapview(chu_geo, zcol = "NAZEV")

prj_noplace <- setdiff(dt$prj_id, dtl$prj_id)

dt_chkonp$prj_id %in% dt$prj_id %>% table()

dt_chkonp %>% 
  # jen projekty v našich datech
  filter(!prj_id %in% dt$prj_id) %>% 
  # jen programu 1-9
  filter(str_detect(prj_id, "CZ\\.0[1-8]")) %>% 
  distinct(prj_id)

prj_noplace %in% dt_chkonp$prj_id %>% table()

dt %>% 
  filter(prj_id %in% prj_noplace) %>% 
  distinct(prj_id, op_id) %>% 
  count(op_id)

dt_chkonp %>%
  left_join(prj_meta) %>% 
  distinct(prj_id, op_id) %>% 
  mutate(noobec = prj_id %in% prj_no_obec$prj_id, 
         noplace = prj_id %in% prj_noplace) %>% 
  count(op_id, noobec, noplace) %>% 
  spread(noobec, n)

dt_chkonp %>%
  left_join(prj_meta) %>% 
  distinct(prj_id, op_id) %>% 
  mutate(noobec = prj_id %in% prj_no_obec$prj_id, 
         noplace = prj_id %in% prj_noplace) %>% 
  filter(noobec) %>% 
  distinct(prj_id, op_id)

dtl %>% 
  filter(typ == "id") %>% 
  distinct(prj_id, prj_nazev, p_nazev, level) %>% 
  filter(prj_id %in% prj_no_obec$prj_id & prj_id %in% dt_chkonp$prj_id) %>% 
  left_join(dt_geostructure_bylvl)

dtl %>% 
  filter(prj_id %in% dt_chkonp$prj_id) %>% 
  group_by(prj_id) %>% 
  mutate(toplevel = min(as.numeric(level))) %>% 
  group_by(op_id, prj_id, level, toplevel, p_nazev, prj_nazev) %>% 
  summarise(n_units = length(unique(value)))

dir.create(here::here("data-processed", "dtl_resolved_placeless"))

dt_with_chuobce_reshaped <- dt_with_chuobce %>% 
  mutate(level = if_else(zuj_id %in% ids_short$obec, "obec", "zuj"), 
         level_orig = "chkonp") %>% 
  rename_with(~str_remove(., "zuj_")) %>% 
  rename(id_orig = chkonp_idtext, value = id) %>% 
  mutate(obec_puvod = "doplnění obce nebo ZÚJ podle chráněných území",
         rozpad_typ = "nic",
         level = factor(level, levels = geolevels, ordered = T),
         rozpad_duvod = NA_character_,
         id_orig = NA_character_,
         level_orig = "chkonp") %>% 
  left_join(prj_meta %>% select(prj_id, op_id)) %>% 
  group_by(op_id) %>% 
  mutate(chunk = floor(row_number()/5e5) + 1,
         chunk = as.integer(chunk)) %>% 
  group_by(prj_id) %>% 
  mutate(radek = row_number()) %>% 
  group_by(obec_puvod, op_id, chunk, rozpad_typ)

write_parquet(dt_with_chuobce_reshaped %>% 
                select(prj_id, radek, op_id, level, geo_id = value, level_orig,
                       geo_id_orig = id_orig,
                       rozpad_typ, rozpad_duvod, obec_puvod, chunk),
              here::here("data-processed", "dtl_resolved_placeless", "all.parquet"))
write_parquet(dt_with_chuobce_reshaped, 
              here::here("data-processed", "dt_noplace_filled-chudetail.parquet"))