编辑文件 - OFXFile.php

✏️ 正在编辑: OFXFile.php

路径: /srv/systems_dir/yuppiecred/library/Core/OFXFile.php
提示: 您可以编辑任何文件（包括二进制文件），但请注意不当修改可能导致文件损坏。

<?php

/**
 * @file
 * Captures the contents of an OFX file
 * 
 * Oct 8 2012			Giving up on a more efficient parse algorithm for now. Implementing more functionality around getting the data
 * Sept 26 2012:  This code is still in flux. I'm currently concentrating on a nifty pair of RegExp's to parse the thing along with recursion.
 *                The problem is currently in finding single line parameters between two aggregates
 */
define('OFXFILEREPORTLEVEL_ERROR', '1');
define('OFXFILEREPORTLEVEL_INFO', '3');
define('OFXFILEREPORTLEVEL_VERBOSE', '5');
define('OFXFILEREPORTLEVEL_DEBUG', '9');
define('OFXFILEREPORTLEVEL_DEV', '10');

/**
 * OFXFile
 * Parses and represents a standard OFX file
 *
 * TODO add DTD validation
 * TODO move reporting out of this class
 */
class Core_OFXFile
{

protected $_data;

/**
     * Construct
     */
    public function __construct($filepath, $reportlevel)
    {
        $this->_data = array();
        $this->_data['infile'] = $filepath;                // file path to the ofx file
        $this->_data['rawheader'] = false;                 // contents of the file's raw header (ofx 1.0)
        $this->_data['reportlevel'] = $reportlevel;        // sets verbosity of reporting
        $this->_data['isDescribed'] = false;               // used to flag discovery of ofx file description (header type)
        $this->_data['doc'] = array();                     // the parsed ofx file DOM
        // preg_match can be limited on the number of backtracks it can perform. For smaller source text, 
        // this rarely becomes a problem. But when applying a regexp to an entire document, you can run into
        // this limit and cause unpredictable results.
        if( !ini_set('pcre.backtrack_limit', 1000000) ){
            $this->report("Couldn't change pcre backtrack limit. If file is large, this may not work...", OFXFILEREPORTLEVEL_ERROR);
            throw new Exception("Couldn't change pcre backtrack limit.");
        }
    }

/**
     * Magic method - get
     */
    public function __get($property)
    {
        if( isset($this->_data[$property]) ){
            return $this->_data[$property];
        }
    }

/**
     * Magic method - set
     */
    public function __set($property, $value)
    {
        $this->_data[$property] = $value;
    }

/**
     * parse the ofx file
     * 
     * If all goes right, the OFX file will be parsed and sitting in the doc variable
     * @return void
     */
    public function parse()
    {
        $linecount = 0;

$file_as_string = file_get_contents($this->_data['infile']);

if( $file_as_string ){
            $file_as_string = trim($file_as_string);

$matches = array();
            if( preg_match('/(?s)((?:[\w\d\:\s]*){9,9})?(<OFX>.*<\/OFX>)/', $file_as_string, $matches) ){
                if( count($matches > 2) ){
                    // must be version < 200
                    $this->rawheader = $matches[1];

if( preg_match_all('/(.*):(.*)/', $this->rawheader, $linematches) ){
                        for ( $i = 0; $i < count($linematches[1]); $i++ ) {
                            $this->_data[$linematches[1][$i]] = $linematches[2][$i];
                        }
                    }
                    $this->isDescribed = true;
                }

if( !$this->isDescribed ){
                    // must be version > 200
                    // TODO get the version info out of the attributes
                }

$this->_data['aggregateList'] = array();
                $this->getNode($matches[2], $this->_data['doc'], $this->_data['aggregateList'], 0);

// TODO this works for MY bank for a Checking account. What happens if this is a Credit Card acct? Or some other type of acct?
                // Looking at the OFX DTD 1.03, I'm not so sure. I think it depends on the bank.
                $this->transactions = $this->doc['OFX'][0]['BANKMSGSRSV1'][0]['STMTTRNRS'][0]['STMTRS'][0]['BANKTRANLIST'][0]['STMTTRN'];

$this->report("Here's the document:", OFXFILEREPORTLEVEL_DEV);
                $this->report(print_r($this->doc, 1), OFXFILEREPORTLEVEL_DEV);
                $this->report("Here's the transactions:", OFXFILEREPORTLEVEL_DEBUG);
                $this->report(print_r($this->getTransactions(), 1), OFXFILEREPORTLEVEL_DEBUG);
            } else {
                $this->report("It's not an OFX file...?", OFXFILEREPORTLEVEL_INFO);
            }
        } else {
            $this->report("Couldn't open file " . $this->infile, OFXFILEREPORTLEVEL_ERROR);
        }
    }

/**
     * query transactions list
     *
     * TODO implement date filter
     */
    public function getTransactions($startDate = false, $endDate = false)
    {
        return $this->transactions;
    }

/**
     * get transactions as a csv string
     *
     * NOTE: according to the DTD the field list in a transaction is variable, i.e., each bank may do it differently
     */
    public function getTransactionsCSV($startDate = false, $endDate = false, $inclHeader = true)
    {
        if( empty($this->csv) ){
            $this->csv = "";
            $txns = $this->getTransactions($startDate, $endDate);

if( !empty($txns) ){
                if( $inclHeader ){
                    $fieldnames = array_keys($txns[0]);
                    $this->csv .= implode(';', $fieldnames) . "\n";
                }

foreach ( $txns as $t ) {
                    $this->csv .= implode(';', $t) . "\n";
                }
            }
        }

return $this->csv;
    }

/**
     * write to stdout
     */
    protected function report($msg, $level)
    {
        if( $level <= $this->reportlevel ){
            print $msg . "\n";
        }
    }

/**
     * recurse into aggregates to discover nested aggregates and elements
     * 
     * This function doesn't use the dtd. It instead traverses the DOM tree discovering aggregates (i.e., tags
     * that are also containers) and elements (i.e., tags that have a single value).
     * 
     * TODO - trying to surf through an OFX SGML file without having to open and parse the OFX DTD
     *      - also trying to use regexp over parsing line by line
     *      - I've run into some problems when trying to discover single line elements.
     *        When getNode receives some sgml, all of the tags are siblings, both aggregates and elements. I only want single line elements (i.e.,
     *        that don't have a closing tag) and I only want single line elements that are at this tree level (i.e., not nested in an aggregate).
     *        I think I have the winning rexexp pattern but php does not grok variable width look behinds, which is what I'm using to discover
     *        opening tags for aggregates (line 171).
     * @param string $sgml          a section of raw OFX SGML
     * @param array $domNode       the current DOM node being parsed
     * @param array $aggregateList  a list of all the aggregate tags (used for discovery)
     * @param int $treeLevel        a counter used to track the current depth of traversal 
     * @param string $nodename      tag name of current DOM Node
     * @return void
     */
    protected function getNode($sgml, &$domNode, &$aggregateList = array(), $treelevel = 0, $nodename = '')
    {
        $this->report("Entering treelevel " . $treelevel . ': ' . $nodename, OFXFILEREPORTLEVEL_DEBUG);

$pattern = ''; // pattern to find aggregates
        $elpattern = ''; // pattern to find elements (single line)

switch ( $treelevel ) {
            case 0:
                $pattern = '/^(?s)(?:\s)*?<([^>]+)>(?>(.*)<\/\1>)(?:\s)*?$/';
                break;
            default:
                $pattern = '/(?s)(?:\s)*?(?><([^>]+)>)(.*?)<\/\1>/';
                /* $elpattern = '/<([^>\/]+)>(.*?)(?:\s)(?>.*(?!(?:<\/[^>]+>)))/'; */
                /* $elpattern = '/(?s)(?<!<[^>]+>\n)<([^>\/]+)>(?!\s)(.*?)(?>(?:\n)(?!.*?<\/[^>]+>))/'; // The first look behind is not fixed-width and php doesn't support variable width look behinds */
                /* $elpattern = '/(?s)<([^>]+)>\n<([^>\/]+)>(?!\s)(?>(.*?)(?:\n).*?)(?!<\/[^>]+>)/'; // */
                $elpattern = '/(?s)(?<!<STMTTRN>)<([^>]+)>(?!\s)(.*?)(?:\n)(?>.*?(?!<\/STMTTRN>))/'; // */
                break;
        }

$this->report("Preg matching...", OFXFILEREPORTLEVEL_DEV);

/**
         * Discovering Aggregates
         *
         * Aggregates can either be singular or multiple. For example, there is only on <OFX> tag, but there are many <STMTTRN> (transaction) tags
         * Up front (without the dtd), we can't distinguish which aggregate tags are multiples. So, as we discover a new tag name for a node, we give it
         * an array to hold all of the instances of that aggregate tag in that node. Most tags only appear once. The result is a multi-dimensional array
         * wherein most arrays only hold one item.  This seems awkward to me, but necessary.
         * 
         */
        if( preg_match_all($pattern, $sgml, $matches) ){

$this->checkPregErrors(); // in doing this, I discovered that php preg has limits on recursion and backtracking; without first setting the proper limits, results are unpredictable

$currentDomNode = null;
            for ( $i = 0; $i < count($matches[1]); $i++ ) {
                $aggregateList[$matches[1][$i]] = true;

if( !isset($domNode[$matches[1][$i]]) ){
                    $domNode[$matches[1][$i]] = array();
                }
                $currentDomNode = &$domNode[$matches[1][$i]];
                $index = count($currentDomNode);
                $currentDomNode[] = array();
                $this->report("Recursing into " . $matches[1][$i], OFXFILEREPORTLEVEL_DEBUG);
                $this->getNode($matches[2][$i], $currentDomNode[$index], $aggregateList, $treelevel + 1, $matches[1][$i]);
            }
        } else {
            $this->report("No aggregates found", OFXFILEREPORTLEVEL_DEBUG);
        }

// Disovering Elements...
        /**
         * This way seems so inefficient. It will work, but I'd like to apply one rexexp to the whole sgml
         * that will find all the parts (see code commented out below)
         * 
         * 20121002 - Checks against a list of all known aggregates. Inefficient, but effective.
         */
        $sgml_lines = preg_split("/((\r?\n)|(\r\n?))/", $sgml);
        $this->report("Looking for elements...", OFXFILEREPORTLEVEL_DEBUG);

$in_aggregate = false;
        foreach ( $sgml_lines as $line ) {
            $line = trim($line);
            if( !empty($line) ){
                if( preg_match('/[\s]*<([^\/>]+)>(.*)/', $line, $elmatches) ){

if( array_key_exists($elmatches[1], $aggregateList) ){
                        $this->report('in aggregate: ' . $elmatches[1], OFXFILEREPORTLEVEL_DEV);
                        $in_aggregate = true;
                    } elseif( $in_aggregate == false ){
                        $this->report("Got an element: " . $elmatches[1] . ' = ' . $elmatches[2], OFXFILEREPORTLEVEL_DEBUG);
                        $domNode[$elmatches[1]] = $elmatches[2];
                    }
                } elseif( preg_match('/<\/[^>]+>/', $line) ){
                    $in_aggregate = false;
                    $this->report('leaving aggregate', OFXFILEREPORTLEVEL_DEV);
                } else {
                    $this->report("Unexpected pattern: " . $line, OFXFILEREPORTLEVEL_DEBUG);
                }
            }
        }

/*         * ******************************************************************
         * This method for discovering elements uses regexp on the whole
         * sgml string. This is a puzzle I don't think I can solve with 
         * php regexp

if($treelevel > 0 && preg_match_all($elpattern,$sgml,$elmatches)) {
          $this->checkPregErrors();
          $this->report('Got some elements', OFXFILEREPORTLEVEL_DEBUG);
          if($treelevel > 2) {
          $this->report(print_r($elmatches,1),OFXFILEREPORTLEVEL_DEBUG);
          }
          } else {
          $this->report("No elements found", OFXFILEREPORTLEVEL_INFO);
          }
          // **************************************************************** */

$this->report("Exiting treelevel " . $treelevel . ': ' . $nodename, OFXFILEREPORTLEVEL_DEBUG);
    }

/**
     * use php's preg error handling to determine success or failure of a preg_match.
     * Reports success and failure back to the user.
     *
     * @return bool
     */
    public function checkPregErrors()
    {
        $result = true;
        switch ( preg_last_error() ) {
            case PREG_INTERNAL_ERROR:
                $this->report("\tPreg internal error", OFXFILEREPORTLEVEL_ERROR);
                $result = false;
                break;
            case PREG_BACKTRACK_LIMIT_ERROR:
                $this->report("\tPreg backtrack limit error", OFXFILEREPORTLEVEL_ERROR);
                $result = false;
                break;
            case PREG_RECURSION_LIMIT_ERROR:
                $this->report("\tPreg recursion limit error", OFXFILEREPORTLEVEL_ERROR);
                $result = false;
                break;
            case PREG_BAD_UTF8_ERROR:
                $this->report("\tPreg bad utf8 error", OFXFILEREPORTLEVEL_ERROR);
                $result = false;
                break;
            case PREG_BAD_UTF8_OFFSET_ERROR:
                $this->report("\tPreg bad utf8 offset error", OFXFILEREPORTLEVEL_ERROR);
                $result = false;
                break;
            default:
                $this->report("\tNo preg errors", OFXFILEREPORTLEVEL_DEV);
        }

return $result;
    }

}

← 返回文件管理器