---
tags: tekstiaine
---

# Tunni jooksul tehtud märkmed (SVUH.00.093 tekstid2)

Koondan siia tunni jooksul tehtud märkmed kergemaks kopeerimiseks

Siia saab koodi kirjutada

tekst_sonad <- raamat1_sonad %>% 
  count(word,sort=T) %>% 
  mutate(sagedus=n/sum(n))

asukohad <- raamat1_sonad %>% 
  mutate(nr=row_number(), n=n()) %>% 
  mutate(asukoht=nr/(n+1)) %>% 
  ungroup()

 
"meie oleme siin" %>% str_detect("me")
"meie oleme siin" %>% str_detect("te")

"meie oleme siin" %>% str_extract("me")
"meie oleme siin" %>% str_extract("te")
"meie oleme siin" %>% str_extract_all("me")

 

 "meie oleme siin" %>% str_extract("^me")
"meie oleme siin" %>% str_extract("\\bme")
"meie oleme siin" %>% str_extract("me\\b")
"meie oleme siin" %>% str_extract("mei\\b")

## Näidispildid


## Isetehtud pildid.

Lisa ctrl+v, siis ta laeb kopeeritud pildi üles ja näitab seda teistelegi



## Rahade kohta

Kolmesõnaline otsing

rahad <- raamat1 %>% 
  mutate(leiud=str_extract_all(txt,"[a-zõäüöA-ZÕÄÖÜ]+ [a-zõäüöA-ZÕÄÖÜ]+ rubla"))

#Meil tekivad tabelisse loendid. Me saame neid lahutada käsuga unnest() valides kõigepealt ainult leidude tulba. 
rahad %>% 
  select(leiud) %>% 
  unnest()

bigrammidega

bigrammid %>% 
  filter(str_detect(bigram,"rubla$"))

Viisgrammide kuvamine

top5grams <- viisgrammid %>% 
  count(fivegram,sort=T) %>% 
  filter(n>3)

viisgrammid %>% 
  mutate(row_number=row_number(),n=n()) %>% 
  mutate(asukoht=row_number/(n+1)) %>% 
  group_by(fivegram) %>% 
  mutate(mitu=n()) %>% 
  filter(mitu>3) %>% 
  ggplot(aes(x=asukoht,y=fivegram)) +
  geom_point()

## Stopsõnad

datadoi link https://datadoi.ee/handle/33/78


Suurtähtede kõrvalejätmiseks täisloendist.

raamat1_sonad2 %>% 
  group_by(chapter) %>% 
  count(word,sort=T) %>% 
  anti_join(stopwords,"word") %>% 
  filter(!str_detect(word,"[A-ZÕÄÖÜ]")) %>%
  mutate(row_number=row_number()) %>% 
  filter(row_number<11) %>% 
  filter(chapter<11) %>% 
  ggplot(aes(x=chapter,y=row_number,label=word))+
  geom_label()

## tf-idf


Peatükid 1-10

Peatükid 11-20

## Mitu teksti korraga

texts %>% 
  filter(str_detect(filename,"Tammsaare")) %>% 
  count(filename)

texts %>% 
  count(filename) %>% 
  ggplot(aes(y=filename,x=n))+
  geom_point()

words %>% 
  count(filename)

words %>% 
  count(word,sort=T)

words %>% 
  count(word,sort=T) %>% 
  mutate(row_number=row_number()) %>%  
  filter(row_number<11) %>% 
  ggplot(aes(x=word,y=n))+
  geom_col()
  
  
words %>% 
  group_by(filename) %>% 
  count(word,sort=T)


words %>% 
  #filter(str_detect(filename,"Ennosaare")) %>% 
  group_by(filename) %>% 
  count(word,sort=T) %>% 
  filter(str_detect(filename,"Ennosaare")) %>% 
  filter(row_number()<11)
  

ja igast tekstist eraldi

words %>% 
  group_by(filename) %>% 
  count(word,sort=T) %>% 
  filter(row_number()<11)

## Veel graafikuid.

words %>% 
  group_by(filename) %>% 
  count(word,sort=T) %>% 
  mutate(freq=n/sum(n))%>% 
  filter(word=="hobune") %>%  #sort pani need juba õigessejärjekorda
  ggplot(aes(y=filename,x=freq,color=filename))+
  geom_point()+
  guides(color=F)
  
words %>% 
  group_by(filename) %>% 
  count(word,sort=T) %>% 
  mutate(freq=n/sum(n))%>% 
  filter(word=="hobune") %>%  #sort pani need juba õigessejärjekorda
  ggplot(aes(y=word,x=freq,color=filename))+
  geom_point()+
  guides(color=F)


words %>% 
  group_by(filename) %>% 
  count(word,sort=T)  %>% 
  mutate(freq=n/sum(n)) %>% 
  filter(str_detect(word,"^töö")) %>%
  summarise(freq=sum(freq)) %>% # tekstid on grupeeritud failinime kaupa
  ggplot(aes(y=filename,x=freq))+
  geom_point()





## Eristavad sõnad

keywords <- texts %>% 
  unnest_tokens(word,txt,to_lower=F) %>% 
  mutate(word=str_replace_all(word,"w","v")) %>% 
  #anti_join(stopwords,"word") %>% 
  group_by(filename) %>% 
  filter(!str_detect(word,"[A-ZÕÄÖÜ]")) %>% 
  count(word,sort=T) %>% 
  bind_tf_idf(word,filename,n) %>% 
  arrange(desc(tf_idf))# %>% 
  #head(100)

keywords %>% 
  filter(str_detect(filename,"vabrik")) 
  

  
keywords <- texts %>% 
  unnest_tokens(word,txt,to_lower=F) %>% 
  mutate(word=str_replace_all(word,"w","v")) %>% 
  #anti_join(stopwords,"word") %>% 
  group_by(filename) %>% 
  filter(!str_detect(word,"[A-ZÕÄÖÜ]")) %>% 
  count(word,sort=T) %>% 
  bind_tf_idf(word,filename,n) %>% 
  arrange(desc(tf_idf)) %>% 
  filter(str_detect(filename,"vabrik")) %>% 
  head(100)

tabel <- words %>% 
  group_by(filename) %>% 
  count(word,sort=T)  %>% 
  mutate(freq=n/sum(n)) %>% 
  mutate(searchterm=str_extract(word,"^mets|^maa|^linn")) %>% 
  filter(!is.na(searchterm)) %>%  # kasutame ainult ridu, kus on olemas info searchterm tulbas
  group_by(filename,searchterm) %>% 
  summarise(freq=sum(freq))

tabel %>% 
  write_tsv("output/mets_maa_linn.tsv")

write_tsv(tabel, "output/mets_maa_linn.tsv")