data import, using RSM strings.
[CommonLispStat.git] / src / data / import.lisp
bloba9b6d6b43e28ee255f091d9b789a02baedf2c70c
1 ;;; -*- mode: lisp -*-
2 ;;; Copyright (c) 2008, by A.J. Rossini <blindglobe@gmail.com>
3 ;;; See COPYRIGHT file for any additional restrictions (BSD license).
4 ;;; Since 1991, ANSI was finally finished. Edited for ANSI Common Lisp.
6 ;;; Time-stamp: <2009-04-20 19:07:46 tony>
7 ;;; Creation: <2008-09-03 08:10:00 tony>
8 ;;; File: import.lisp
9 ;;; Author: AJ Rossini <blindglobe@gmail.com>
10 ;;; Copyright: (c)2007, AJ Rossini. BSD, LLGPL, or GPLv2, depending
11 ;;; on how it arrives.
12 ;;; Purpose: base structures for importing data into CLS
14 ;;; What is this talk of 'release'? Klingons do not make software
15 ;;; 'releases'. Our software 'escapes', leaving a bloody trail of
16 ;;; designers and quality assurance people in its wake.
18 (in-package :cls-dataimport)
25 ;;; Data I/O
27 ;; We can read 2 types of data -- those which are non-lisp-native
28 ;; data, and those which are lisp-native (lisp-enabled, an extension
29 ;; of lisp-serialized, i.e. data as program as data thingy's).
31 ;; of the non-native, there could be raw sources (ascii file formats),
32 ;; xml sources (xml -> lisp, possible with some preprocessing.
34 ;;; Reading from DSV files:
36 ;;; consider either the cybertyggyr-dsv package, or the rsm.string
37 ;;; package. The latter seems to actually work a bit at what we need
38 ;;; to acccomplish, but the former is a challenge to get right when we
39 ;;; need to think abut what it is that we need to get done. The
40 ;;; latter is also better licensed. i.e. BSD-style.
43 (defparameter *lisp-stat-data-external-source-formats*
44 '(csv tsv xml ;; ex of text-based (UTF, ASCII, or similar) formats
45 sql ;; ex of RDBMS call
46 fcs affy)) ;; ex of binary formats
48 (defparameter *lisp-stat-data-import-referencing-type*
49 '(lisp-data-structure reference lisp-function))
51 (defgeneric data-import (source source-format referencing-type)
52 (:documentation "read data from stream srce, in format srce-frmt;
53 return a reftype, which could be a
54 lisp-data-structure, a reference to such, or a lisp
55 function which can be evaluated to generate
56 either."))
58 (defgeneric data-export (data target-format target-referencing-type)
59 (:documentation "write data from stream srce, in format srce-frmt;
60 return a reftype, which could be a
61 lisp-data-structure, a reference to such, or a lisp
62 function which can be evaluated to generate
63 either."))
68 ;;; Potentially useful functions
70 ;; the following belongs here if we are working externally, but might
71 ;; belong with data if we are working internlly
73 ;; (defmacro with-data (body)
74 ;; "Stream-handling, maintaining I/O through object typing.")
77 ;;;
78 ;;; Related to data file reading
79 ;;;
81 (defun count-file-columns (fname)
82 "Args: (fname)
83 Returns the number of lisp items on the first nonblank line of file FNAME."
84 (with-open-file (f fname)
85 (if f
86 (let ((line (do ((line (read-line f) (read-line f)))
87 ((or (null line) (< 0 (length line))) line))))
88 (if line
89 (with-input-from-string (s line)
90 (do ((n 0 (+ n 1)) (eof (gensym)))
91 ((eq eof (read s nil eof)) n))))))))
93 (if (not (fboundp 'open-file-dialog))
94 #+dialogs
95 (defun open-file-dialog () ;; why?(&optional set)
96 (get-string-dialog "Enter a data file name:"))
97 #-dialogs
98 (defun open-file-dialog () ;; why? (&optional set)
99 (error "You must provide a file name explicitly")))
101 (defun read-data-file (&optional (file (open-file-dialog t)))
102 "Args: (file)
103 Returns a list of all lisp objects in FILE. FILE can be a string or a symbol,
104 in which case the symbol'f print name is used."
105 (if file
106 (let ((eof (gensym)))
107 (with-open-file (f file)
108 (if f
109 (do* ((r (read f nil eof) (read f nil eof))
110 (x (list nil))
111 (tail x (cdr tail)))
112 ((eq r eof) (cdr x))
113 (setf (cdr tail) (list r))))))))
115 ;;; New definition to avoid stack size limit in apply
116 (defun read-data-columns (&optional (file (open-file-dialog t))
117 (cols (if file
118 (count-file-columns file))))
119 "Args: (&optional file cols)
120 Reads the data in FILE as COLS columns and returns a list of lists representing the columns."
121 (if (and file cols)
122 (transpose (split-list (read-data-file file) cols))))
125 ;;; FIXME:AJR: ALL THE FOLLOWING NEED TO BE SOLVED BY PLATFORM-INDEP PATHNAME WORK!
126 ;;; FIXME:AJR: use either string or pathname.
128 (defun path-string-to-path (p s)
129 (pathname (concatenate 'string (namestring p) s)))
131 (defun load-data (file)
132 "Args: (file) as string
133 Read in data file from the data examples library."
134 (if (load (path-string-to-path *lispstat-data-dir* file))
136 (load (path-string-to-path *lispstat-data-dir* file))))
138 (defun load-example (file)
139 "Args: (file) as string
140 Read in lisp example file from the examples library."
141 (if (load (path-string-to-path *lispstat-examples-dir* file))
143 (load (path-string-to-path *lispstat-examples-dir* file))))
146 ;;; Saving Variables and Functions
149 (defun savevar (vars file)
150 "Args: (vars file-name-root)
151 VARS is a symbol or a list of symbols. FILE-NAME-ROOT is a string (or a symbol
152 whose print name is used) not endinf in .lsp. The VARS and their current values
153 are written to the file FILE-NAME-ROOT.lsp in a form suitable for use with the
154 load command."
155 (with-open-file (f (concatenate 'string (namestring file) ".lsp")
156 :direction :output)
157 (let ((vars (if (consp vars) vars (list vars))))
158 (flet ((save-one (x)
159 (let ((v (symbol-value x)))
160 (if (objectp v)
161 (format f "(def ~s ~s)~%" x (send v :save))
162 (format f "(def ~s '~s)~%" x v)))))
163 (mapcar #'save-one vars))
164 vars)))
168 ;;; General modification approaches.
170 (defgeneric importData (source featureList)
171 (:documentation "command to get data into CLS. Specific methods
172 will need to handle pathnames, internal data structures, and
173 external services such as DBMS's. We would like to be able to do
174 thinks like:
175 (importData MyPathName '(:formattype 'csvString))
176 (importData '(sqlConnection :server host.domain.net :port 666)
177 '(:formattype 'table
178 and so on."))
181 (defun pathname-example (name)
182 (let ((my-path (parse-namestring name)))
183 (values (pathname-name my-path :case :common)
184 (pathname-name my-path :case :local))))
186 (defvar sourceTypes (list 'csv 'lisp 'tsv 'special)
187 "list of possible symbols.
189 Thsees are used to specify source formats that might be supported for
190 input. CSV and TSV are standard, LISP refers to forms, and SPECIAL
191 refers to a FUNCTION which parses as appropriately.")
193 ;;; WRONG LOGIC.
194 (defmethod importData ((fileHandle pathname)
195 (fmt list)) ;sourceTypes))
196 "File-based input for data.
197 Usually used by:
198 (importData (parse-namestring 'path/to/file')
199 (list :format 'csv))
201 (importData myPathName (list :format 'lisp))
203 (let* ((fmtType (getf fmt :format))
204 (newData (getDataAsLists fileHandle fmtType)))
205 (case fmtType
206 ('csv ( ))
207 ('tsv ( ))
208 ('lisp ( ))
209 ('special (let ((parserFcn (getf fmt :special-parser)))))
210 (:default (error "no standard default importData format")))))
212 (defmethod importData ((ds array) (fmt list))
213 "mapping arrays into CLS data.")
215 (defmethod importData ((dsSpec DBMSandSQLextract)
216 (fmt mappingTypes))
217 "mapping DBMS into CLS data.")
221 ;;(defmacro with-dataframe (env &rest progn)
222 ;; "Compute using variable names with with.data.frame type semantics.")