library( irs990efile ) # functions for parsing efile XMLs
library( dplyr ) # data wrangling
library( purrr ) # data wrangling
library( pander ) # formatting
library( knitr ) # formatting
library( DT ) # table printing
library( httr ) # web requests
library( xmltools ) # xml utilities
library( xml2 ) # xml utilities
library( XML ) # xml utilities
library( data.tree ) # network visualization
library( networkD3 ) # network visualization
library( igraph ) # network visualization
source( "R/rdb-functions-v2.R")
source( "R/utils.R")
head( index )
The concordance is an RDB to XML Crosswalk that maps all xpaths onto a specified set of variables. It is located in the irs990efile package.
The full data dictionary is available on GitHub.
%>%
concordance filter( rdb_table == table.name )
xpath | id |
---|---|
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/BondReferenceCd | 1 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/CUSIPNumber | 2 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/DateIssued | 3 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/Defeased | 4 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/DescriptionOfPurpose | 5 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuePrice | 6 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuerEIN | 7 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuerName/BusinessNameLine1 | 8 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuerName/BusinessNameLine2 | 9 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/OnBehalfOfIssuer | 10 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/PoolFinancing | 11 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/BondIssuedDt | 12 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/BondIssuerEIN | 13 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/BondReferenceCd | 14 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/CUSIPNum | 15 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/DefeasedInd | 16 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuePriceAmt | 17 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine1 | 18 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine1Txt | 19 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine2 | 20 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine2Txt | 21 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/OnBehalfOfIssuerInd | 22 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/PoolFinancingInd | 23 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/PurposeDesc | 24 |
<- as.character( t.xpaths$xpath )
xpaths <- create_edgelist_v1( xpaths )
el <- FromDataFrameNetwork( network=el )
nd print( nd )
## levelName
## 1 Return
## 2 °--ReturnData
## 3 °--IRS990ScheduleK
## 4 ¦--Form990ScheduleKPartI
## 5 ¦ ¦--BondReferenceCd
## 6 ¦ ¦--CUSIPNumber
## 7 ¦ ¦--DateIssued
## 8 ¦ ¦--Defeased
## 9 ¦ ¦--DescriptionOfPurpose
## 10 ¦ ¦--IssuePrice
## 11 ¦ ¦--IssuerEIN
## 12 ¦ ¦--IssuerName
## 13 ¦ ¦ ¦--BusinessNameLine1
## 14 ¦ ¦ ¦--BusinessNameLine2
## 15 ¦ ¦ ¦--BusinessNameLine1Txt
## 16 ¦ ¦ °--BusinessNameLine2Txt
## 17 ¦ ¦--OnBehalfOfIssuer
## 18 ¦ °--PoolFinancing
## 19 °--TaxExemptBondsIssuesGrp
## 20 ¦--BondIssuedDt
## 21 ¦--BondIssuerEIN
## 22 ¦--BondReferenceCd
## 23 ¦--CUSIPNum
## 24 ¦--DefeasedInd
## 25 ¦--IssuePriceAmt
## 26 ¦--IssuerName
## 27 ¦ ¦--BusinessNameLine1
## 28 ¦ ¦--BusinessNameLine2
## 29 ¦ ¦--BusinessNameLine1Txt
## 30 ¦ °--BusinessNameLine2Txt
## 31 ¦--OnBehalfOfIssuerInd
## 32 ¦--PoolFinancingInd
## 33 °--PurposeDesc
SetGraphStyle( nd, rankdir = "LR")
SetEdgeStyle( nd, arrowhead = "vee", color = "grey20", penwidth = 2 )
SetNodeStyle( nd,
style = "filled,rounded",
shape = "box",
fillcolor = "LightBlue",
fontname = "helvetica",
fontcolor="black",
tooltip = GetDefaultTooltip )
# SetNodeStyle(acme$IT, fillcolor = "LightBlue", penwidth = "5px")
plot( nd )
# old version
find_group_names( table.name )
## [1] "//Form990ScheduleKPartI" "//TaxExemptBondsIssuesGrp"
# new version
find_table_headers( table.name )
## [1] "//Form990ScheduleKPartI/BondReferenceCd"
## [2] "//Form990ScheduleKPartI/IssuerName"
## [3] "//Form990ScheduleKPartI/IssuerEIN"
## [4] "//Form990ScheduleKPartI/CUSIPNumber"
## [5] "//Form990ScheduleKPartI/DateIssued"
## [6] "//Form990ScheduleKPartI/IssuePrice"
## [7] "//Form990ScheduleKPartI/DescriptionOfPurpose"
## [8] "//Form990ScheduleKPartI/Defeased"
## [9] "//Form990ScheduleKPartI/OnBehalfOfIssuer"
## [10] "//Form990ScheduleKPartI/PoolFinancing"
## [11] "//TaxExemptBondsIssuesGrp/BondReferenceCd"
## [12] "//TaxExemptBondsIssuesGrp/IssuerName"
## [13] "//TaxExemptBondsIssuesGrp/BondIssuerEIN"
## [14] "//TaxExemptBondsIssuesGrp/CUSIPNum"
## [15] "//TaxExemptBondsIssuesGrp/BondIssuedDt"
## [16] "//TaxExemptBondsIssuesGrp/IssuePriceAmt"
## [17] "//TaxExemptBondsIssuesGrp/PurposeDesc"
## [18] "//TaxExemptBondsIssuesGrp/DefeasedInd"
## [19] "//TaxExemptBondsIssuesGrp/OnBehalfOfIssuerInd"
## [20] "//TaxExemptBondsIssuesGrp/PoolFinancingInd"
# manual
if( ! is.null(table.headers) )
{table.headers}
# create new folder
dir.create( folder.name )
setwd( folder.name )
# create sample
set.seed( 1234 )
<- sample_n( index, sample.size )
sample.index write.csv( index, "sample-index.csv", row.names=F )
<- sample.index$URL
sample.urls
# erase existing log files
file.create("XPATH-LOG.txt")
file.create("FAIL-LOG.txt")
<- Sys.time() # --------------------
start.build.time
<- list()
results.list <- list()
xpath.list
for( i in 1:length( sample.urls ) )
{
<- sample.urls[i]
url
<-
results.list[[i]] build_rdb_table_v2( url,
table.name, table.headers=table.headers )
<-
xpath.list[[i]] get_table_xpaths( url,
table.name, table.headers=table.headers )
if( i %% 100 == 0 ){ print(i) }
}
<- Sys.time() # --------------------
end.build.time
<- dplyr::bind_rows( results.list )
df
<- readLines( "FAIL-LOG.txt" )
fail.log <- xpath.list %>% unlist() %>% unique() %>% sort()
all.xpaths cat( all.xpaths, sep="\n", file="XPATH-LOG.txt" )
TABLE: SK-P01-T01-BOND-ISSUES
## [1] "NUMBER OF RETURNS PROCESSED: 1000"
## [1] "TOTAL NUMBER OF TABLE ROWS: 2087"
TOTAL RUN TIME:
## Time difference of 30.81794 mins
## Time difference of -0.5136323 hours
write.csv( df, paste0( folder.name, "/", table.name, ".csv" ), row.names=F )
head(df)
Are all names standardized?
names(df)
## [1] "OBJECT_ID" "EIN"
## [3] "NAME" "TAXYR"
## [5] "FORMTYPE" "URL"
## [7] "SK_01_BOND_ISSUE_REFERENCE_NUM" "SK_01_BOND_ISSUER_EIN"
## [9] "SK_01_BOND_ISSUE_DATE" "SK_01_BOND_ISSUE_PRICE"
## [11] "SK_01_BOND_ISSUE_PURPOSE_DESC" "SK_01_BOND_ISSUE_DEFEASED_X"
## [13] "SK_01_BOND_ISSUER_ON_BEHALF_OF_X" "SK_01_BOND_ISSUE_POOL_FIN_X"
## [15] "SK_01_BOND_ISSUER_NAME_L1" "SK_01_BOND_ISSUE_CUSIP_NUM"
## [17] "SK_01_BOND_ISSUER_NAME_L2"
Check for xpaths that were used in the tables but do not exist in the Concordance file.
<- dplyr::filter( concordance, rdb_table == table.name )
c.table <- c.table$xpath
xpath.concordance setdiff( all.xpaths, xpath.concordance )
## [1] ""
## [2] "/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName"
Capture to update the Concordance file.
<- setdiff( all.xpaths, xpath.concordance )
these cat( these, sep="\n", file=paste0( folder.name, "/MISSING-XPATHS.txt" ) )
Check the status of failed URLs.
if( length(fail.log) > 0 )
{ sapply( fail.log, get_message ) %>%
table() %>%
pander()
else {
} print("ALL URLs ARE VALID")
}
## [1] "ALL URLs ARE VALID"
The create_edgelist() functions v1 and v2 are in R/utils.R.
V1 retains the root nodes Return and ReturnData.
V2 drops the root nodes and uses form versions “990” and “990EZ” as the root to visualize the sets of tables separately.
# KEEP ROOT NOTES Return & ReturnData
<- create_edgelist_v1( all.xpaths ) el
Edgelist format:
V1 | V2 |
---|---|
Return | ReturnData |
ReturnData | IRS990 |
IRS990 | Form990PartVIISectionAGrp |
Form990PartVIISectionAGrp | PersonNm |
Form990PartVIISectionAGrp | TitleTxt |
Form990PartVIISectionAGrp | AverageHoursPerWeekRt |
Form990PartVIISectionAGrp | IndividualTrusteeOrDirectorInd |
Form990PartVIISectionAGrp | OfficerInd |
Form990PartVIISectionAGrp | ReportableCompFromOrgAmt |
Form990PartVIISectionAGrp | ReportableCompFromRltdOrgAmt |
# library( data.tree )
<- FromDataFrameNetwork( network=el )
nd print( nd )
## levelName
## 1 Return
## 2 °--ReturnData
## 3 °--IRS990ScheduleK
## 4 °--TaxExemptBondsIssuesGrp
## 5 ¦--BondIssuedDt
## 6 ¦--BondIssuerEIN
## 7 ¦--BondReferenceCd
## 8 ¦--CUSIPNum
## 9 ¦--DefeasedInd
## 10 ¦--IssuePriceAmt
## 11 ¦--IssuerName
## 12 ¦ ¦--BusinessNameLine1Txt
## 13 ¦ °--BusinessNameLine2Txt
## 14 ¦--OnBehalfOfIssuerInd
## 15 ¦--PoolFinancingInd
## 16 °--PurposeDesc
SetGraphStyle( nd, rankdir = "LR")
SetEdgeStyle( nd, arrowhead = "vee", color = "grey20", penwidth = 2 )
SetNodeStyle( nd,
style = "filled,rounded",
shape = "box",
fillcolor = "LightBlue",
fontname = "helvetica",
fontcolor="black",
tooltip = GetDefaultTooltip )
# SetNodeStyle(acme$IT, fillcolor = "LightBlue", penwidth = "5px")
plot( nd )
# library( networkD3 )
# net <- ToDataFrameNetwork( nd, "name" )
# simpleNetwork( net, fontSize = 12 )
simpleNetwork( el, fontSize = 12 )
#plot with networkD3
<- ToListExplicit( nd, unname = TRUE )
radnet radialNetwork( radnet, fontSize = 12 )
<-
t.xpaths %>%
concordance filter( rdb_table == table.name ) %>%
select( xpath ) %>%
unique() %>%
arrange( xpath ) %>%
mutate( id=row_number() )
%>%
t.xpaths kable( align="l" )
xpath | id |
---|---|
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/BondReferenceCd | 1 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/CUSIPNumber | 2 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/DateIssued | 3 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/Defeased | 4 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/DescriptionOfPurpose | 5 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuePrice | 6 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuerEIN | 7 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuerName/BusinessNameLine1 | 8 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/IssuerName/BusinessNameLine2 | 9 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/OnBehalfOfIssuer | 10 |
/Return/ReturnData/IRS990ScheduleK/Form990ScheduleKPartI/PoolFinancing | 11 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/BondIssuedDt | 12 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/BondIssuerEIN | 13 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/BondReferenceCd | 14 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/CUSIPNum | 15 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/DefeasedInd | 16 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuePriceAmt | 17 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine1 | 18 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine1Txt | 19 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine2 | 20 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/IssuerName/BusinessNameLine2Txt | 21 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/OnBehalfOfIssuerInd | 22 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/PoolFinancingInd | 23 |
/Return/ReturnData/IRS990ScheduleK/TaxExemptBondsIssuesGrp/PurposeDesc | 24 |
Identify all unique groups used in xpaths.
if( is.null(table.headers) )
find_table_headers( table.name ) } else
{ table.headers
## [1] "//Form990ScheduleKPartI/BondReferenceCd"
## [2] "//Form990ScheduleKPartI/IssuerName"
## [3] "//Form990ScheduleKPartI/IssuerEIN"
## [4] "//Form990ScheduleKPartI/CUSIPNumber"
## [5] "//Form990ScheduleKPartI/DateIssued"
## [6] "//Form990ScheduleKPartI/IssuePrice"
## [7] "//Form990ScheduleKPartI/DescriptionOfPurpose"
## [8] "//Form990ScheduleKPartI/Defeased"
## [9] "//Form990ScheduleKPartI/OnBehalfOfIssuer"
## [10] "//Form990ScheduleKPartI/PoolFinancing"
## [11] "//TaxExemptBondsIssuesGrp/BondReferenceCd"
## [12] "//TaxExemptBondsIssuesGrp/IssuerName"
## [13] "//TaxExemptBondsIssuesGrp/BondIssuerEIN"
## [14] "//TaxExemptBondsIssuesGrp/CUSIPNum"
## [15] "//TaxExemptBondsIssuesGrp/BondIssuedDt"
## [16] "//TaxExemptBondsIssuesGrp/IssuePriceAmt"
## [17] "//TaxExemptBondsIssuesGrp/PurposeDesc"
## [18] "//TaxExemptBondsIssuesGrp/DefeasedInd"
## [19] "//TaxExemptBondsIssuesGrp/OnBehalfOfIssuerInd"
## [20] "//TaxExemptBondsIssuesGrp/PoolFinancingInd"
Groups can then be used to extract the 1:M RDB tables:
if( is.null(table.headers) )
<- find_table_headers( table.name ) }
{ table.headers <- "https://nccs-efile.s3.us-east-1.amazonaws.com/xml/201913079349300206_public.xml"
demo.url <- xml2::read_xml( demo.url )
doc ::xml_ns_strip( doc )
xml2<- get_table_v2( doc, table.name, table.headers )
df.demo df.demo
## NULL
Mapping of xml variable names to concordance variable names:
get_var_map( table.name=table.name )
A table with proper names:
<- get_var_map( table.name=table.name )
v.map
if( ! is.null(df.demo) )
re_name( df.demo, v.map ) %>% as.data.frame()} {
Demo of functions - demo data only.
[[1]]
<Form990PartVIISectionAGrp>
<PersonNm>
{text}
<TitleTxt>
{text}
<AverageHoursPerWeekRt>
{text}
<IndividualTrusteeOrDirectorInd>
{text}
<OfficerInd>
{text}
<ReportableCompFromOrgAmt>
{text}
<ReportableCompFromRltdOrgAmt>
{text}
<OtherCompensationAmt>
{text}
1]] %>% xml_parent() %>% xmltools::xml_view_trees() node.list[[
(1) ------------
+-- Form990PartVIISectionAGrp
+-- PersonNm
+-- TitleTxt
+-- AverageHoursPerWeekRt
+-- IndividualTrusteeOrDirectorInd
+-- OfficerInd
+-- ReportableCompFromOrgAmt
+-- ReportableCompFromRltdOrgAmt
++-- OtherCompensationAmt
(2) ------------
+-- Form990PartVIISectionAGrp
+-- PersonNm
+-- TitleTxt
+-- AverageHoursPerWeekRt
+-- IndividualTrusteeOrDirectorInd
+-- OfficerInd
+-- ReportableCompFromOrgAmt
+-- ReportableCompFromRltdOrgAmt
++-- OtherCompensationAmt