Mercurial > repos > jjohnson > query_tabular
view query_tabular.py @ 16:5b4f6cf857cf draft
Uploaded
author | jjohnson |
---|---|
date | Thu, 22 Jun 2017 09:14:56 -0400 |
parents | fd16243931d6 |
children | b9f797bf4f38 |
line wrap: on
line source
#!/usr/bin/env python """ """ import sys import re import os.path import json import sqlite3 as sqlite import optparse from optparse import OptionParser """ TODO: - could read column names from comment lines, but issues with legal names - could add some transformations on tabular columns, filter - skip_regex e.g. a regex to format date/time strings format: { c2 : re.sub('pat', 'sub', c2) c3 : len(c3) } def format(colname,val, expr): normalize input list columns iterate over list values creating one row per iteration option for input line_num column create associated table fk, name, value # e.g. PSM table with list of proteins containing peptide fk, name, value[, value] # if multiple columns similarly indexed, e.g. vcf - column_defs dict of columns to create from tabular input column_defs : { 'name1' : 'expr', 'name2' : 'expr'} - allow multiple queries and outputs repeat min - max with up to max conditional outputs - add a --json input for table definitions (or yaml) JSON config: { tables : [ { file_path : '/home/galaxy/dataset_101.dat', table_name : 't1', column_names : ['c1', 'c2', 'c3'], pkey_autoincr : 'id' comment_lines : 1 unique: ['c1'], index: ['c2','c3'] }, { file_path : '/home/galaxy/dataset_102.dat', table_name : 'gff', column_names : ['seqname',,'date','start','end'] comment_lines : 1 load_named_columns : True filters : [{'filter': 'regex', 'pattern': '#peptide', 'action': 'exclude_match'}, {'filter': 'replace', 'column': 3, 'replace': 'gi[|]', 'pattern': ''}] }, { file_path : '/home/galaxy/dataset_103.dat', table_name : 'test', column_names : ['c1', 'c2', 'c3'] } ] } """ class LineFilter( object ): def __init__(self,source,filter_dict): self.source = source self.filter_dict = filter_dict # print >> sys.stderr, 'LineFilter %s' % filter_dict if filter_dict else 'NONE' self.func = lambda i,l: l.rstrip('\r\n') if l else None self.src_lines = [] self.src_line_cnt = 0 if not filter_dict: return if filter_dict['filter'] == 'regex': rgx = re.compile(filter_dict['pattern']) if filter_dict['action'] == 'exclude_match': self.func = lambda i,l: l if not rgx.match(l) else None elif filter_dict['action'] == 'include_match': self.func = lambda i,l: l if rgx.match(l) else None elif filter_dict['action'] == 'exclude_find': self.func = lambda i,l: l if not rgx.search(l) else None elif filter_dict['action'] == 'include_find': self.func = lambda i,l: l if rgx.search(l) else None elif filter_dict['filter'] == 'select_columns': cols = [int(c) - 1 for c in filter_dict['columns']] self.func = lambda i,l: self.select_columns(l,cols) elif filter_dict['filter'] == 'replace': p = filter_dict['pattern'] r = filter_dict['replace'] c = int(filter_dict['column']) - 1 self.func = lambda i,l: '\t'.join([x if i != c else re.sub(p,r,x) for i,x in enumerate(l.split('\t'))]) elif filter_dict['filter'] == 'prepend_line_num': self.func = lambda i,l: '%d\t%s' % (i,l) elif filter_dict['filter'] == 'append_line_num': self.func = lambda i,l: '%s\t%d' % (l.rstrip('\r\n'),i) elif filter_dict['filter'] == 'skip': cnt = filter_dict['count'] self.func = lambda i,l: l if i > cnt else None elif filter_dict['filter'] == 'normalize': cols = [int(c) - 1 for c in filter_dict['columns']] sep = filter_dict['separator'] self.func = lambda i,l: self.normalize(l,cols,sep) def __iter__(self): return self def select_columns(self,line,cols): fields = line.split('\t') return '\t'.join([fields[x] for x in cols]) def normalize(self,line,split_cols,sep): lines = [] fields = line.rstrip('\r\n').split('\t') split_fields = dict() cnt = 0 for c in split_cols: if c < len(fields): split_fields[c] = fields[c].split(sep) cnt = max(cnt, len(split_fields[c])) if cnt == 0: lines.append('\t'.join(fields)) else: for n in range(0, cnt): flds = [x if c not in split_cols else split_fields[c][n] if n < len(split_fields[c]) else '' for (c, x) in enumerate(fields)] lines.append('\t'.join(flds)) return lines def get_lines(self): for i,next_line in enumerate(self.source): self.src_line_cnt += 1 line = self.func(self.src_line_cnt,next_line) # print >> sys.stderr, 'LineFilter %s: %d %s' % (str(self.filter_dict),self.src_line_cnt,line) if line: if isinstance(line,list): self.src_lines.extend(line) else: self.src_lines.append(line) return def next(self): if not self.src_lines: self.get_lines() if self.src_lines: return self.src_lines.pop(0) raise StopIteration class TabularReader: """ Tabular file iterator. Returns a list """ def __init__(self, file_path, skip=0, comment_char=None, col_idx=None, filters=None): self.skip = skip self.comment_char = comment_char self.col_idx = col_idx self.filters = filters self.tsv_file = open(file_path) if skip and skip > 0: for i in range(skip): if not self.tsv_file.readline(): break source = LineFilter(self.tsv_file,None) if comment_char: source = LineFilter(source,{"filter": "regex", "pattern": comment_char, "action": "exclude_match"}) if filters: for f in filters: source = LineFilter(source,f) self.source = source def __iter__(self): return self def next(self): ''' Iteration ''' for i,line in enumerate(self.source): fields = line.rstrip('\r\n').split('\t') if self.col_idx: fields = [fields[i] for i in self.col_idx] return fields raise StopIteration def getValueType(val): if val or 0. == val: try: int(val) return 'INTEGER' except: try: float(val) return 'REAL' except: return 'TEXT' return None def get_column_def(file_path, table_name, skip=0, comment_char='#', column_names=None, max_lines=100,load_named_columns=False,filters=None): col_pref = ['TEXT', 'REAL', 'INTEGER', None] col_types = [] col_idx = None data_lines = 0 try: tr = TabularReader(file_path,skip=skip, comment_char=comment_char, col_idx=None, filters=filters) for linenum, fields in enumerate(tr): if linenum > max_lines: break try: while len(col_types) < len(fields): col_types.append(None) for i, val in enumerate(fields): colType = getValueType(val) if col_pref.index(colType) < col_pref.index(col_types[i]): col_types[i] = colType except Exception, e: print >> sys.stderr, 'Failed at line: %d err: %s' % (linenum, e) except Exception, e: print >> sys.stderr, 'Failed: %s' % (e) for i,col_type in enumerate(col_types): if not col_type: col_types[i] = 'TEXT' if column_names: col_names = [] if load_named_columns: col_idx = [] for i, cname in enumerate([cn.strip() for cn in column_names.split(',')]): if cname != '': col_idx.append(i) col_names.append(cname) col_types = [col_types[i] for i in col_idx] else: col_names = ['c%d' % i for i in range(1, len(col_types) + 1)] for i, cname in enumerate([cn.strip() for cn in column_names.split(',')]): if cname and i < len(col_names): col_names[i] = cname else: col_names = ['c%d' % i for i in range(1, len(col_types) + 1)] col_def = [] for i, col_name in enumerate(col_names): col_def.append('%s %s' % (col_names[i], col_types[i])) return col_names, col_types, col_def, col_idx def create_table(conn, file_path, table_name, skip=0, comment_char='#', pkey_autoincr=None, column_names=None,load_named_columns=False,filters=None,unique_indexes=[],indexes=[]): col_names, col_types, col_def, col_idx = get_column_def(file_path, table_name, skip=skip, comment_char=comment_char, column_names=column_names,load_named_columns=load_named_columns,filters=filters) col_func = [float if t == 'REAL' else int if t == 'INTEGER' else str for t in col_types] table_def = 'CREATE TABLE %s (\n %s%s\n);' % ( table_name, '%s INTEGER PRIMARY KEY AUTOINCREMENT,' % pkey_autoincr if pkey_autoincr else '', ', \n '.join(col_def)) # print >> sys.stdout, table_def insert_stmt = 'INSERT INTO %s(%s) VALUES(%s)' % (table_name, ','.join(col_names), ','.join(["?" for x in col_names])) # print >> sys.stdout, insert_stmt data_lines = 0 try: c = conn.cursor() c.execute(table_def) conn.commit() c.close() for i,index in enumerate(unique_indexes): index_name='idx_uniq_%s_%d' % (table_name,i) index_columns = index.split(',') create_index(conn, table_name, index_name, index_columns,unique=True) for i,index in enumerate(indexes): index_name='idx_%s_%d' % (table_name,i) index_columns = index.split(',') create_index(conn, table_name, index_name, index_columns) c = conn.cursor() tr = TabularReader(file_path,skip=skip, comment_char=comment_char, col_idx=col_idx, filters=filters) for linenum, fields in enumerate(tr): data_lines += 1 try: vals = [col_func[i](x) if x else None for i, x in enumerate(fields)] c.execute(insert_stmt, vals) except Exception, e: print >> sys.stderr, 'Failed at line: %d err: %s' % (linenum, e) conn.commit() c.close() except Exception, e: print >> sys.stderr, 'Failed: %s' % (e) exit(1) def create_index(conn, table_name, index_name, index_columns, unique=False): index_def = "CREATE %s INDEX %s on %s(%s)" % ('UNIQUE' if unique else '', index_name, table_name, ','.join(index_columns)) c = conn.cursor() c.execute(index_def) conn.commit() c.close() def regex_match(expr, item): return re.match(expr, item) is not None def regex_search(expr, item): return re.search(expr, item) is not None def regex_sub(expr, replace, item): return re.sub(expr, replace, item) def get_connection(sqlitedb_path, addfunctions=False): conn = sqlite.connect(sqlitedb_path) if addfunctions: conn.create_function("re_match", 2, regex_match) conn.create_function("re_search", 2, regex_search) conn.create_function("re_sub", 3, regex_sub) return conn def __main__(): # Parse Command Line parser = optparse.OptionParser() parser.add_option('-s', '--sqlitedb', dest='sqlitedb', default=None, help='The SQLite Database') parser.add_option('-t', '--table', dest='tables', action="append", default=[], help='Tabular file: file_path[=table_name[:column_name, ...]') parser.add_option('-j', '--jsonfile', dest='jsonfile', default=None, help='Tabular file: file_path[=table_name[:column_name, ...]') parser.add_option('-q', '--query', dest='query', default=None, help='SQL query') parser.add_option('-Q', '--query_file', dest='query_file', default=None, help='SQL query file') parser.add_option('-n', '--no_header', dest='no_header', action='store_true', default=False, help='Include a column headers line') parser.add_option('-o', '--output', dest='output', default=None, help='Output file for query results') (options, args) = parser.parse_args() def run_query(query,outputFile): conn = get_connection(options.sqlitedb, addfunctions=True) cur = conn.cursor() results = cur.execute(query) if not options.no_header: outputFile.write("#%s\n" % '\t'.join([str(col[0]) for col in cur.description])) # yield [col[0] for col in cur.description] for i, row in enumerate(results): # yield [val for val in row] outputFile.write("%s\n" % '\t'.join([str(val) if val is not None else '' for val in row])) # open sqlite connection conn = get_connection(options.sqlitedb) # determine output destination if options.output is not None: try: outputPath = os.path.abspath(options.output) outputFile = open(outputPath, 'w') except Exception, e: print >> sys.stderr, "failed: %s" % e exit(3) else: outputFile = sys.stdout # get table defs if options.tables: for ti, table in enumerate(options.tables): table_name = 't%d' % (ti + 1) column_names = None fields = table.split('=') path = fields[0] if len(fields) > 1: names = fields[1].split(':') table_name = names[0] if names[0] else table_name if len(names) > 1: column_names = names[1] # print >> sys.stdout, '%s %s' % (table_name, path) create_table(conn, path, table_name, column_names=column_names) if options.jsonfile: try: fh = open(options.jsonfile) tdef = json.load(fh) if 'tables' in tdef: for ti, table in enumerate(tdef['tables']): path = table['file_path'] table_name = table['table_name'] if 'table_name' in table else 't%d' % (ti + 1) comment_lines = table['comment_lines'] if 'comment_lines' in table else 0 comment_char = table['comment_char'] if 'comment_char' in table else None column_names = table['column_names'] if 'column_names' in table else None if column_names: load_named_columns = table['load_named_columns'] if 'load_named_columns' in table else False else: load_named_columns = False unique_indexes = table['unique'] if 'unique' in table else [] indexes = table['index'] if 'index' in table else [] filters = table['filters'] if 'filters' in table else None pkey_autoincr = table['pkey_autoincr'] if 'pkey_autoincr' in table else None create_table(conn, path, table_name, pkey_autoincr=pkey_autoincr, column_names=column_names, skip=comment_lines, comment_char=comment_char, load_named_columns=load_named_columns, filters=filters,unique_indexes=unique_indexes, indexes=indexes) except Exception, exc: print >> sys.stderr, "Error: %s" % exc conn.close() query = None if (options.query_file is not None): with open(options.query_file, 'r') as fh: query = '' for line in fh: query += line elif (options.query is not None): query = options.query if (query is None): tables_query = \ "SELECT name, sql FROM sqlite_master WHERE type='table' ORDER BY name" try: conn = get_connection(options.sqlitedb) c = conn.cursor() rslt = c.execute(tables_query).fetchall() for table, sql in rslt: print >> sys.stderr, "Table %s:" % table try: col_query = 'SELECT * FROM %s LIMIT 0' % table cur = conn.cursor().execute(col_query) cols = [col[0] for col in cur.description] print >> sys.stderr, " Columns: %s" % cols except Exception, exc: print >> sys.stderr, "Error: %s" % exc except Exception, exc: print >> sys.stderr, "Error: %s" % exc exit(0) # if not sqlite.is_read_only_query(query): # print >> sys.stderr, "Error: Must be a read only query" # exit(2) try: run_query(query,outputFile) except Exception, exc: print >> sys.stderr, "Error: %s" % exc exit(1) if __name__ == "__main__": __main__()