Seitenhistorie

#!/bin/bash

set -euo pipefail
# set -x

# https://wiki.dnb.de/displaypages/GND/1.+Mappings+GND+-+Fremdthesauri,
# https://wiki.dnb.de/display/GND/5.+Datenformat+GND+-+Crosskonkordanzen+zu+externen+Thesauri
viewpage.action?pageId=263851158

dnb_dump=/srv/aen-data/pica/T.dat

# Tc-Sätze "f bbg Tc NOT rdb GND-kein-Schlagwort*"
pica filter -s "002@.0 == 'Tc'" --not "050C.a =^ 'GND-kein-Schlagwort'" $dnb_dump -o Tc.dat

# 028P (700) - "p" - Person
# 029P (710) - "b" - Körperschaft
# 030P (711) - "f" - Konferenz
# 022P (730) - "u" - Einheitstitel
# 041P (750) - "s" - Sachbegriff
# 065P (751) - "g" - Geografikum

pica filter -s "028P.4?" Tc.dat |
pica select -H "IDN, Feld, Thesaurus, Relation" \
"003@.0, '028P', 028P{2, 4}" \
--where "028P.4?" Tc.dat -o Tc.csv

# ohne Header anhängen
pica filter -s "029P.4?" Tc.dat |
pica select "003@.0, '029P', 029P{2, 4}" >> Tc.csv

pica filter -s--where "030P029P.4?" Tc.dat | --append -o Tc.csv
pica select "003@.0, '030P', 030P{2, 4}" >> Tc.csv

pica filter -s--where "022P030P.4?" Tc.dat | --append -o Tc.csv
pica select "003@.0, '022P', 022P{2, 4}" >> Tc.csv

pica filter -s--where "041P022P.4?" Tc.dat | --append -o Tc.csv
pica select "003@.0, '041P', 041P{2, 4}" >> Tc.csv

pica filter -s--where "065P041P.4?" Tc.dat | --append -o Tc.csv
pica select "003@.0, '065P', 065P{2, 4}" >>--where "065P.4?" Tc.dat --append -o Tc.csv

# Crosskonkordanzen einlesen
ck <- read_csv('Tc.csv', col_types = 'cccc')

# Wertebereiche gemäß Wiki-Seiten definieren
thesauri <- read_csv('Thesauri.csv', col_types = 'cc')
relationen <- read_csv('Relationen.csv', col_types = 'cc')

# Crosskonkordanzen filtern und zählen
rslt1 <- ck %>%
  filter(Thesaurus %in% thesauri$Code & Relation %in% relationen$Code) %>%
  mutate(
    Thesaurus = factor(Thesaurus, 
                       levels = thesauri$Code, labels = thesauri$Thesaurus),
    Relation  = factor(Relation,
                       levels = relationen$Code, labels = relationen$Relation)
  ) %>% # auch alle leeren Levels jeweils mit Anzahl 0 anzeigen
  group_by(Thesaurus, Relation, .drop = FALSE) %>% 
  # jede Kombination (Thesaurus, Relation) je IDN nur einmal zählen
  summarise(n = n_distinct(IDN)) %>%
  pivot_wider(id_cols = Relation, names_from = Thesaurus, values_from = n)

# Summe
rslt2 <- data.frame(Relation = factor('GESAMT'), 
                    rslt1 %>% select(-Relation) %>% summarise_all(.funs = sum))
colnames(rslt2) <- colnames(rslt1)

# Formatierte Ausgabetabelle
rbind(rslt1, rslt2) %>% 
  mutate_at(.vars = setdiff(colnames(.), 'Relation'), .funs = ~ formatC(
    ., format = 'd', big.mark = '.', decimal.mark = ',')) %>%
  knitr::kable(align = c('l', rep('r', ncol(.)-1)))