This file contains the code necessary for parsing Plato’s Republic from the Diorisis Ancient Greek Corpus (Vatri, Alessandro and Barbara McGillivray. 2018. ‘The Diorisis Ancient Greek Corpus’, Research Data Journal for the Humanities and Social Sciences 3, 55–65). With this code, each book is saved as a separate file; to parse the whole text, please refer to our Corpus_Platonicum (Diorisis): Lemmata Extraction. I here assume that you have already have the “Republic.xml” file in your working directory; if not, please refer to the link above.
First load the file from the working directory
library(XML)
doc_R <- xmlParse("Republic.xml", useInternalNodes = TRUE, isURL = F)
rootnode_R <- xmlRoot(doc_R)Here’s how the xml looks like. The number of books is stored under the “location” attribute in “sentence”; each sentence contains a number of words, and lemmata are stored under “lemma” element (we need the attribute “entry” in Greek).
<body>
<sentence id="1" location="1.327">
<word form="kate/bhn" id="1">
<lemma id="53651" entry="καταβαίνω" POS="verb" TreeTagger="false" disambiguated="n/a">
<analysis morph="aor ind act 3rd pl (epic doric aeolic)"/>
<analysis morph="aor ind act 1st sg"/>
</lemma>
</word>
<word form="xqe\s" id="2">
<lemma id="113946" entry="χθές" POS="adverb" TreeTagger="false" disambiguated="n/a">
<analysis morph="indeclform (adverb)"/>
</lemma>
</word>We need to subset sentences according to the book they belong to.
library(stringr)
all_sent <- getNodeSet(rootnode_R, "//text/body/sentence") ## 6389 elements
b1 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 1
b2 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 2
b3 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 3
b4 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 4
b5 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 5
b6 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 6
b7 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 7
b8 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 8
b9 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 9
b10 <- str_extract((sapply(all_sent, xmlGetAttr, 'location')), "\\d+(?=\\.)") == 10
sent_b1 <- all_sent[b1]
sent_b2 <- all_sent[b2]
sent_b3 <- all_sent[b3]
sent_b4 <- all_sent[b4]
sent_b5 <- all_sent[b5]
sent_b6 <- all_sent[b6]
sent_b7 <- all_sent[b7]
sent_b8 <- all_sent[b8]
sent_b9 <- all_sent[b9]
sent_b10 <- all_sent[b10]Before we extract the information stored in under “lemma” elements in each sentence, we need to know the number of sentences in each book.
lb1 <- length(sent_b1)
lb2 <- length(sent_b2)
lb3 <- length(sent_b3)
lb4 <- length(sent_b4)
lb5 <- length(sent_b5)
lb6 <- length(sent_b6)
lb7 <- length(sent_b7)
lb8 <- length(sent_b8)
lb9 <- length(sent_b9)
lb10 <- length(sent_b10)Book 1 is the longest (847 sentences), book 9 is the shortest (525 sentences). There are 6389 sentences in the “Republic”.
c(lb1, lb2, lb3, lb4, lb5, lb6, lb7, lb8, lb9, lb10)## [1] 847 607 670 637 817 578 572 571 525 565We first want to write a loop that would go through all the sentences and extracts all the lemmata for book 1.
text_b1 <- character()
for (n in 1:lb1) {
words <- sent_b1[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b1 <- c(text_b1, lemma)
}
}
length(text_b1)## [1] 9469text_b1 <- as.character(text_b1)
head(text_b1)## [1] "καταβαίνω" "χθές" "εἰς" "Πειραιεύς" "μετά" "Γλαύκων"write.table(text_b1, "Republic1.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b2 <- character()
for (n in 1:lb2) {
words <- sent_b2[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b2 <- c(text_b2, lemma)
}
}
length(text_b2)## [1] 8920text_b2 <- as.character(text_b2)
head(text_b2)## [1] "ἐγώ" "μέν" "οὖν" "οὗτος" "εἶπον" "οἴομαι"write.table(text_b2, "Republic2.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b3 <- character()
for (n in 1:lb3) {
words <- sent_b3[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b3 <- c(text_b3, lemma)
}
}
length(text_b3)## [1] 10276text_b3 <- as.character(text_b3)
head(text_b3)## [1] "καί" "οἴομαι" "γε" "φημί" "ὀρθός" "ἐγώ"write.table(text_b3, "Republic3.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)Note that 386a1-4 ( Τὰ μὲν δὴ περὶ θεούς, ἦν δ’ ἐγώ, τοιαῦτ’ ἄττα, ὡς ἔοικεν, ἀκουστέον τε καὶ οὐκ ἀκουστέον εὐθὺς ἐκ παίδων τοῖς θεούςτε τιμήσουσιν καὶ γονέας τήν τε ἀλλήλων φιλίαν μὴ περὶ σμικροῦ ποιησομένοις.)is parsed as the last sentence in book 2. Book 3 in this file starts with 386a5 (Καὶ οἶμαί γ’, ἔφη, ὀρθῶς ἡμῖν φαίνεσθαι.)
text_b4 <- character()
for (n in 1:lb4) {
words <- sent_b4[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b4 <- c(text_b4, lemma)
}
}
length(text_b4)## [1] 8898text_b4 <- as.character(text_b4)
head(text_b4)## [1] "καί" "ὁ" "Ἀδείμαντος" "ὑπολαμβάνω" "τίς"
## [6] "οὖν"write.table(text_b4, "Republic4.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b5 <- character()
for (n in 1:lb5) {
words <- sent_b5[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b5 <- c(text_b5, lemma)
}
}
length(text_b5)## [1] 10239text_b5 <- as.character(text_b5)
head(text_b5)## [1] "ἀγαθός" "μέν" "τοίνυν" "ὁ" "τοιοῦτος" "πόλις"write.table(text_b5, "Republic5.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b6 <- character()
for (n in 1:lb6) {
words <- sent_b6[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b6 <- c(text_b6, lemma)
}
}
length(text_b6)## [1] 8718text_b6 <- as.character(text_b6)
head(text_b6)## [1] "ἕ" "μέν" "δή" "φιλόσοφος" "εἰμί" "δέ"write.table(text_b6, "Republic6.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b7 <- character()
for (n in 1:lb7) {
words <- sent_b7[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b7 <- c(text_b7, lemma)
}
}
length(text_b7)## [1] 8592text_b7 <- as.character(text_b7)
head(text_b7)## [1] "μετά" "οὗτος" "δή" "εἶπον" "ἀπεικάζω" "τοιοῦτος"write.table(text_b7, "Republic7.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b8 <- character()
for (n in 1:lb8) {
words <- sent_b8[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b8 <- c(text_b8, lemma)
}
}
length(text_b8)## [1] 8195text_b8 <- as.character(text_b8)
head(text_b8)## [1] "εἰμί" "οὗτος" "μέν" "δή" "ὁμολογέω" "ὦ"write.table(text_b8, "Republic8.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b9 <- character()
for (n in 1:lb9) {
words <- sent_b9[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b9 <- c(text_b9, lemma)
}
}
length(text_b9)## [1] 6858text_b9 <- as.character(text_b9)
head(text_b9)## [1] "αὐτός" "δή" "λοιπός" "εἰμί" "δέ" "ἐγώ"write.table(text_b9, "Republic9.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)text_b10 <- character()
for (n in 1:lb10) {
words <- sent_b10[[n]]["word"]
sl <- length(words)
for (y in 1:sl) {
lemma <- sapply(words[[y]]["lemma"], xmlGetAttr, "entry")
text_b10 <- c(text_b10, lemma)
}
}
length(text_b10)## [1] 8713text_b10 <- as.character(text_b10)
head(text_b10)## [1] "καί" "μήν" "εἰμί" "δέ" "ἐγώ" "πολύς"write.table(text_b10, "Republic10.txt", row.names = FALSE, col.names = FALSE, quote = FALSE)We now have all books of the “Laws” saved as txt files in the working directory.
list.files(pattern = "^Rep")## [1] "Republic.txt" "Republic.xml" "Republic1.txt" "Republic10.txt"
## [5] "Republic2.txt" "Republic3.txt" "Republic4.txt" "Republic5.txt"
## [9] "Republic6.txt" "Republic7.txt" "Republic8.txt" "Republic9.txt"There is probably a more economic way to solve the problem, but the output is as expected (the head of each book above confirms this).
Please note that the original xml file contains some lemmatization errors, but this problem is to be addressed separately.
With questions, corrections, and suggestions please write to Olga Alieva (oalieva@hse.ru)