This assignment will begin exploring the trade-offs posed by
structured data (= stored in database) and unstructured data (= plain
text). In particular, we will explore the "data cleaning" problem, and
the ease (or difficulty) with which we can extract useful information
from unstructured data and query it.

The file dinesafe.xml (available in the class dataset page) contains
information about a large number of food safety inspections for
Toronto-based establishments that prepare and/or sell food: everything
from hot dog vending carts to restaurants to grocery stores to meat
packing plants. We are particularly interested in the
INFRACTION_DETAILS field: it contains information about cases where
the inspector found operator and/or employees failed to comply with
the various laws governing the handling and sale of food in
Toronto. The dataset was acquired from Toronto's DineSafe Inspection
and Disclosure System [1] and covers inspections over the last three
years. The inspection system makes use of three laws, which are
summarized below:

Ontario Health Protection and Promotion Act, R.S.O 1990, ch. H.7 [2]
----------------------------------------------------------------------

Grants the Ontario Ministry of Health authority to "provide for the
organization and delivery of public health programs and services, the
prevention of the spread of disease and the promotion and protection
of the health of the people of Ontario" (Sec. 3). 


Ontario Regulation 562 - Food Premises [3]
----------------------------------------------------------------------

Lists specific rules which must be followed when preparing and serving
food. It covers buildings and equipment (construction, maintenance,
and cleanliness), employee hygiene, and all aspects of food storage
and preparation. The majority of food safety infractions will cite a
specific section of this law.


Toronto Municipal Code, chapter 545 - Licensing [4]
----------------------------------------------------------------------

Section 5.G of this law requires all eating and drinking
establishments to be properly licensed. The business itself, and all
food handlers working there, must obtain proper licenses and present
them upon request.

[1] http://www.toronto.ca/health/dinesafe
[2] http://www.e-laws.gov.on.ca/html/statutes/english/elaws_statutes_90h07_e.htm
[3] http://www.e-laws.gov.on.ca/html/regs/english/elaws_regs_900562_e.htm
[4] http://www.toronto.ca/legdocs/municode/1184_545.pdf

========================================================================

For this assignment, you should use shell tools (sort, grep, sed, wc,
etc.) and write python code to analyze the data and answer the
following questions:

Q1: How many infractions are recorded in the dataset? How many
    distinct infractions are there?

Q2: Of the distinct infractions found, how many arise out of the
    Health Protection Act vs. the Municipal Code?

Q3: What irregularities in the data made it harder to answer Q1 and
    Q2? How did you work around those irregularities?

Q4: Not all infractions cite any law. Using other information
    (e.g. from other infractions in the dataset and the text of the
    laws [1][2]), how many of those infractions can you match with a
    specific section (subsection, paragraph) of a law? Your code
    should take a sequence of infraction details as input and output
    groups of infractions that fall under the same section of law. For
    example:

    O. Reg. 562/90 Sec. 68(4)
        fail to clean toilets as often as necessary
        fail to clean toilets once a day
        fail to clean washbasins as often as necessary
        fail to clean washbasins once a day
        ...
    O. Reg. 562/90 Sec. 23
        fail to store food on pallets
        fail to store food on racks or shelves
    ...

Q5: Which section of the law has the most infractions? How many?

Q6: (BONUS, CHALLENGING) Of inspections that closed a business, what
    rule was cited the most often? How many times? What detailed
    reasons were given?

Because this is real data, it will be "messy" and potentially
ambiguous. That is part of the "fun." You will find that
case-insensitive comparisons and regular expressions can be extremely
useful.

========================================================================