sql-blocks 0.0.3__py3-none-any.whl → 0.0.5__py3-none-any.whl

sql_blocks/sql_blocks.py

@@ -1,13 +1,28 @@
 from enum import Enum
+import re
+
+
+PATTERN_PREFIX = '([^0-9 ]+[.])'
+PATTERN_SUFFIX = '( [A-Za-z_]+)'
+SUFFIX_AND_PRE = f'{PATTERN_SUFFIX}|{PATTERN_PREFIX}'
+DISTINCT_SF_PR = f'(DISTINCT|distinct)|{SUFFIX_AND_PRE}'
 
 KEYWORD = {
-    'SELECT': (',{}', 'SELECT *'),
-    'FROM': ('{}', ''),
-    'WHERE': ('{}AND ', ''),
-    'GROUP BY': (',{}', ''),
-    'ORDER BY': (',{}', ''),
-    'LIMIT': (' ', ''),
+    'SELECT': (',{}', 'SELECT *', DISTINCT_SF_PR),
+    'FROM': ('{}', '', PATTERN_SUFFIX),
+    'WHERE': ('{}AND ', '', PATTERN_PREFIX),
+    'GROUP BY': (',{}', '', SUFFIX_AND_PRE),
+    'ORDER BY': (',{}', '', SUFFIX_AND_PRE),
+    'LIMIT': (' ', '', ''),
 }
+#             ^     ^        ^
+#             |     |        |
+#             |     |        +----- pattern to compare fields
+#             |     |
+#             |     +----- default when empty (SELECT * ...)
+#             |
+#             +-------- separator
+
 SELECT, FROM, WHERE, GROUP_BY, ORDER_BY, LIMIT = KEYWORD.keys()
 USUAL_KEYS = [SELECT, WHERE, GROUP_BY, ORDER_BY]
 
@@ -37,6 +52,26 @@ class SQLObject:
     def table_name(self) -> str:
         return self.values[FROM][0].split()[0]
  
+    @staticmethod
+    def get_separator(key: str) -> str:
+        appendix = {WHERE: 'and|', FROM: 'join|JOIN'}
+        return KEYWORD[key][0].format(appendix.get(key, ''))
+
+    def diff(self, key: str, search_list: list, symmetrical: bool=False) -> set:
+        pattern = KEYWORD[key][2]
+        separator = self.get_separator(key)
+        def field_set(source: list) -> set:
+            return set(
+                re.sub(pattern, '', fld.strip()).lower()
+                for string in source
+                for fld in re.split(separator, string)
+            )       
+        s1 = field_set(search_list)
+        s2 = field_set(self.values.get(key, []))
+        if symmetrical:
+            return s1.symmetric_difference(s2)
+        return s1 - s2
+
     def delete(self, search: str, keys: list=USUAL_KEYS):
         for key in keys:
             result = []
@@ -54,9 +89,10 @@ class Field:
 
     @classmethod
     def format(cls, name: str, main: SQLObject) -> str:
+        name = name.strip()
         if name == '_':
             name = '*'
-        else:
+        elif '.' not in name:
             name = f'{main.alias}.{name}'
         if Function in cls.__bases__:
             name = f'{cls.__name__}({name})'
@@ -319,20 +355,19 @@ class JoinType(Enum):
 
 class Select(SQLObject):
     join_type: JoinType = JoinType.INNER
-    REGEX = None
+    REGEX = {}
 
     def __init__(self, table_name: str='', **values):
         super().__init__(table_name)
         self.__call__(**values)
         self.break_lines = True
 
+    def update_values(self, key: str, new_values: list):
+        for value in self.diff(key, new_values):
+            self.values.setdefault(key, []).append(value)
+
     def add(self, name: str, main: SQLObject):
-        def update_values(key: str, new_values: list):
-            for value in new_values:
-                old_values = main.values.get(key, [])
-                if value not in old_values:
-                    main.values[key] = old_values + [value]
-        update_values(
+        main.update_values(
             FROM, [
                 '{jt}JOIN {tb} {a2} ON ({a1}.{f1} = {a2}.{f2})'.format(
                     jt=self.join_type.value,
@@ -343,14 +378,18 @@ class Select(SQLObject):
             ] + self.values[FROM][1:]
         )
         for key in USUAL_KEYS:
-            update_values(key, self.values.get(key, []))
+            main.update_values(key, self.values.get(key, []))
 
     def __add__(self, other: SQLObject):
+        if self.table_name.lower() == other.table_name.lower():
+            for key in USUAL_KEYS:
+                self.update_values(key, other.values.get(key, []))
+            return self
         foreign_field, primary_key = ForeignKey.find(self, other)
         if not foreign_field:
             foreign_field, primary_key = ForeignKey.find(other, self)
             if foreign_field:
-                if primary_key and not self.key_field:
+                if primary_key:
                     PrimaryKey.add(primary_key, self)
                 self.add(foreign_field, other)
                 return other
@@ -380,10 +419,8 @@ class Select(SQLObject):
         return self
 
     def __eq__(self, other: SQLObject) -> bool:
-        def sorted_values(obj: SQLObject, key: str) -> list:
-            return sorted(obj.values.get(key, []))
         for key in KEYWORD:
-            if sorted_values(self, key) != sorted_values(other, key):
+            if self.diff(key, other.values.get(key, []), True):
                 return False
         return True
 
@@ -396,15 +433,52 @@ class Select(SQLObject):
 
     @classmethod
     def parse(cls, txt: str) -> list[SQLObject]:
-        import re
-        txt = re.sub(' +', ' ', re.sub('\n|\t', ' ', txt))
+        def find_last_word(pos: int) -> int:
+            SPACE, WORD = 1, 2
+            found = set()
+            for i in range(pos, 0, -1):
+                if txt[i] in [' ', '\t', '\n']:
+                    if sum(found) == 3:
+                        return i
+                    found.add(SPACE)
+                if txt[i].isalpha():
+                    found.add(WORD)
+                elif txt[i] == '.':
+                    found.remove(WORD)
+        def find_parenthesis(pos: int) -> int:
+            for i in range(pos, len(txt)-1):
+                if txt[i] == ')':
+                    return i+1
         if not cls.REGEX:
-            keywords = '|'.join(KEYWORD)
-            cls.REGEX = re.compile(f'({keywords})', re.IGNORECASE)
-        tokens = [t.strip() for t in cls.REGEX.split(txt) if t.strip()]
-        values = {k: v for k, v in zip(tokens[::2], tokens[1::2])}
-        tables = [t.strip() for t in re.split('JOIN|LEFT|RIGHT|ON', values[FROM]) if t.strip()]
+            keywords = '|'.join(k + r'\b' for k in KEYWORD)
+            flags = re.IGNORECASE + re.MULTILINE
+            cls.REGEX['keywords'] = re.compile(f'({keywords})', flags)
+            cls.REGEX['subquery'] = re.compile(r'(\w\.)*\w+ +in +\(SELECT.*?\)', flags)
         result = {}
+        found = cls.REGEX['subquery'].search(txt)
+        while found:
+            start, end = found.span()
+            inner = txt[start: end]
+            if inner.count('(') > inner.count(')'):
+                end = find_parenthesis(end)
+                inner = txt[start: end-1]
+            fld, *inner = re.split(r' IN | in', inner, maxsplit=1)
+            if fld.upper() == 'NOT':
+                pos = find_last_word(start)
+                fld = txt[pos: start].strip() # [To-Do] Use the value of `fld`
+                start = pos
+                class_type = NotSelectIN
+            else:
+                class_type = SelectIN
+            obj = class_type.parse(
+                ' '.join(re.sub(r'^\(', '', s.strip()) for s in inner)
+            )[0]
+            result[obj.alias] = obj
+            txt = txt[:start-1] + txt[end+1:]
+            found = cls.REGEX['subquery'].search(txt)
+        tokens = [t.strip() for t in cls.REGEX['keywords'].split(txt) if re.findall(r'\w+', t)]
+        values = {k.upper(): v for k, v in zip(tokens[::2], tokens[1::2])}
+        tables = [t.strip() for t in re.split('JOIN|LEFT|RIGHT|ON', values[FROM]) if t.strip()]
         for item in tables:
             if '=' in item:
                 a1, f1, a2, f2 = [r.strip() for r in re.split('[().=]', item) if r]
@@ -417,23 +491,52 @@ class Select(SQLObject):
                 for key in USUAL_KEYS:
                     if not key in values:
                         continue
-                    separator = KEYWORD[key][0].format(
-                        ' ' if key == WHERE else ''
-                    )
-                    fields = [
-                        fld.strip() for fld in re.split(
-                            separator, values[key]
-                        ) if len(tables) == 1
-                        or re.findall(f'^[( ]*{obj.alias}.', fld)
-                    ]
+                    separator = cls.get_separator(key)
                     obj.values[key] = [
-                        f'{obj.alias}.{f}' if not '.' in f else f for f in fields
+                        Field.format(fld, obj)
+                        for fld in re.split(separator, values[key])
+                        if len(tables) == 1 or re.findall(f'\b*{obj.alias}[.]', fld)
                     ]
                 result[obj.alias] = obj
         return list( result.values() )
 
+class SelectIN(Select):
+    condition_class = Where
 
-class SubSelect(Select):
     def add(self, name: str, main: SQLObject):
         self.break_lines = False
-        Where.list(self).add(name, main)
+        self.condition_class.list(self).add(name, main)
+
+SubSelect = SelectIN
+
+class NotSelectIN(SelectIN):
+    condition_class = Not
+
+
+if __name__ == "__main__":
+    query_list = Select.parse("""
+        SELECT
+                cas.role,
+                m.title,
+                m.release_date,
+                a.name as actors_name
+        FROM
+                Actor a
+                LEFT JOIN Cast cas ON (a.cast = cas.id)
+                LEFT JOIN Movie m ON (cas.movie = m.id)
+        WHERE
+                m.genre NOT in (SELECT g.id from Genres g where g.name in ('sci-fi', 'horror', 'distopia'))
+                AND (m.hashtag = '#cult' OR m.awards LIKE '%Oscar%')
+                AND m.id IN (select DISTINCT r.movie FROM Review r GROUP BY r.movie HAVING Avg(r.rate) > 4.5)
+                AND a.age <= 69 AND a.age >= 45
+        ORDER BY
+                m.release_date
+    """)
+    for query in query_list:
+        descr = ' {} ({}) '.format(
+            query.table_name,
+            query.__class__.__name__
+        )
+        print(descr.center(50, '-'))
+        print(query)
+    print('='*50)

{sql_blocks-0.0.3.dist-info → sql_blocks-0.0.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sql_blocks
-Version: 0.0.3
+Version: 0.0.5
 Summary: Allows you to create objects for parts of SQL query commands. Also to combine these objects by joining them, adding or removing parts...
 Home-page: https://github.com/julio-cascalles/sql_blocks
 Author: Júlio Cascalles
@@ -55,7 +55,7 @@ You can specify your own alias:  `a = Select('Actor a')`
     3.2 -- Sub-queries:
 ```
 query = Select('Movie m', title=Field,
-    id=SubSelect(
+    id=SelectIN(
         'Review r',
         rate=Where.gt(4.5),
         movie_id=Distinct
@@ -86,12 +86,13 @@ query = Select('Movie m', title=Field,
 
 3.4 -- Negative conditions use the _Not_ class instead of _Where_
 ```
-franchise
 based_on_book=Not.is_null()
 ```
 
 3.5 -- List of values
+```
 hash_tag=Where.list(['space', 'monster', 'gore'])
+```
 
 ---
 ### 4 - A field can be two things at the same time:
@@ -100,7 +101,7 @@ hash_tag=Where.list(['space', 'monster', 'gore'])
     - This means that the field will appear in the results and also that the query will be ordered by that field.
 * Applying **GROUP BY** to item 3.2, it would look like this:
     ```    
-    SubSelect(
+    SelectIN(
         'Review r', movie=[GroupBy, Distinct],
         rate=Having.avg(Where.gt(4.5))
     )
@@ -248,13 +249,13 @@ m = Select...
 9.1
 ```
         a1 = Select.parse('''
-                SELECT gender, max(age) FROM Actor act
+                SELECT gender, Max(act.age) FROM Actor act
                 WHERE act.age <= 69 AND act.age >= 45
                 GROUP BY gender
             ''')[0]
 
         a2 = Select('Actor',
-            age=Between(45, 69), gender=GroupBy,
+            age=[ Between(45, 69), Max ],
             gender=[GroupBy, Field]
         )       
 ```
@@ -282,7 +283,7 @@ m = Select...
 
 9.3
 ```
-best_movies = SubSelect(
+best_movies = SelectIN(
     Review=Table('role'),
     rate=[GroupBy, Having.avg(Where.gt(4.5))]
 )
@@ -302,13 +303,13 @@ m2 = Select(
 ---
 
 ### 10 - CASE...WHEN...THEN
-Select(
-    'Product',
-    label=Case('price').when(
-        lt(50), 'cheap'
-    ).when(
-        gt(100), 'expensive'
-    ).else_value(
-        'normal'
+    Select(
+        'Product',
+        label=Case('price').when(
+            lt(50), 'cheap'
+        ).when(
+            gt(100), 'expensive'
+        ).else_value(
+            'normal'
+        )
     )
-)

sql_blocks-0.0.5.dist-info/RECORD

@@ -0,0 +1,7 @@
+sql_blocks/__init__.py,sha256=TodC5q-UEdYEz9v1RRoogVqqRcsKnZRY1WDGinrI2zo,26
+sql_blocks/sql_blocks.py,sha256=lTdPxelcWQC9oRFGuDcdEm2g5r6q1typyQAc7sGjmVY,17230
+sql_blocks-0.0.5.dist-info/LICENSE,sha256=6kbiFSfobTZ7beWiKnHpN902HgBx-Jzgcme0SvKqhKY,1091
+sql_blocks-0.0.5.dist-info/METADATA,sha256=xi_F_iiC8Wr8Hgk6WTp0tQ_KBiv3rDOpSA5hhVulzdk,7025
+sql_blocks-0.0.5.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+sql_blocks-0.0.5.dist-info/top_level.txt,sha256=57AbUvUjYNy4m1EqDaU3WHeP-uyIAfV0n8GAUp1a1YQ,11
+sql_blocks-0.0.5.dist-info/RECORD,,

sql_blocks-0.0.3.dist-info/RECORD

@@ -1,7 +0,0 @@
-sql_blocks/__init__.py,sha256=TodC5q-UEdYEz9v1RRoogVqqRcsKnZRY1WDGinrI2zo,26
-sql_blocks/sql_blocks.py,sha256=r_PIxS-VqaY2zuBGwBgz6Z1Z2zBrx_CrfZ35yTMb7Sw,13352
-sql_blocks-0.0.3.dist-info/LICENSE,sha256=6kbiFSfobTZ7beWiKnHpN902HgBx-Jzgcme0SvKqhKY,1091
-sql_blocks-0.0.3.dist-info/METADATA,sha256=MHZb5SM_2RCZim4wESLsLaiCYReDOkn7tYJwGNsGWBE,6992
-sql_blocks-0.0.3.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-sql_blocks-0.0.3.dist-info/top_level.txt,sha256=57AbUvUjYNy4m1EqDaU3WHeP-uyIAfV0n8GAUp1a1YQ,11
-sql_blocks-0.0.3.dist-info/RECORD,,