99爱在线视频这里只有精品_窝窝午夜看片成人精品_日韩精品久久久毛片一区二区_亚洲一区二区久久

合肥生活安徽新聞合肥交通合肥房產(chǎn)生活服務(wù)合肥教育合肥招聘合肥旅游文化藝術(shù)合肥美食合肥地圖合肥社保合肥醫(yī)院企業(yè)服務(wù)合肥法律

代寫 COMP3506、代做 Python 語言程序

時間:2024-08-19  來源:合肥網(wǎng)hfw.cc  作者:hfw.cc 我要糾錯



Assignment One – 15%
Algorithms and Data Structures – COMP3506/7505 – Semester 2, 2024
Due: 3pm on Friday August 23 (week 5)
Summary
The main objective of this assignment is to get your hands dirty with some simple data structures
and algorithms to solve basic computational problems. These data structures will also come in handy
for your second assignment, so you should take your time to think about your implementations and
try to make them as efficient as possible.
1 Getting Started
Before we get into the nitty gritty, we will discuss the skeleton codebase that will form
the basis of your implementations, and provide some rules that must be followed when
implementing your solutions.
1.1 Codebase
The codebase contains a number of data structures stubs that you should implement, as well
as some scripts that allow your code to be tested. Figure 1 shows a snapshot of the project
directory tree with the different files categorized.
test_structures.py
test_warmup.py
test_kmers.py
structures
dynamic_array.py
linked_list.py
bit_vector.py
warmup
warmup.py
malloclabs
kmer_structure.py
analysis.txt
generate_dna.py
Figure 1 The directory tree is organized by task. Blue represents directories, Teal represents files
that contain implementations (but are not executable), and Orange represents executable files.
1.2 Implementation Rules
The following list outlines some important information regarding the skeleton code, and your
implementation. If you have any doubts, please ask on Ed discussion.
? The code is written in Python and, in particular, should be executed with Python 3
or higher. The EAIT student server, moss, has Python 3.11.* installed by default. We
recommend using moss for the development and testing of your assignment, but you can
use your own system if you wish.
2 COMP3506/7505 – Semester 2, 2024
? You are not allowed to use built-in methods or data structures – this is an algorithms and
data structures course, after all. If you want to use a dict (aka {}), you will need to imple-
ment that yourself. Lists can be used as “dumb arrays” by manually allocating space like
myArray = [None] * 10 but you may not use built-ins like append, clear, count,
copy, extend, index, insert, pop, remove, reverse, sort, min, max, and so
on. List slicing is also prohibited, as are functions like sorted, len, reversed, zip.
Be sensible – if you need the functionality provided by these methods, you may implement
them yourself. Similarly, don’t use any other collections or structures such as set or
tuple (for example; mytup = ("abc", 123)).
? You are not allowed to use libraries such as numpy, pandas, scipy, collections, and so
on.
? Exceptions: The only additional libraries you can use are random and math. You are
allowed to use range and enumerate to handle looping. You can also use for item in
my_list looping over simple lists.
2 Task 1: Data Structures (5 marks)
We’ll start off by implementing some fundamental data structures. You should write your
own tests. We will try to break your code via (hidden) corner cases. You have been warned.
Task 1.1: Doubly Linked List (1.5 Marks)
Your first task is to implement a doubly linked list — your first “pointer-based” data structure.
To get started, look at the linked_list.py file. You will notice that this file contains two
classes: the Node type, which stores a data payload, as well as a reference to the next node;
and the DoublyLinkedList type which tracks the head and tail of the list, as well as the number
of nodes in the list.
A basic set of functions that you need to support are provided as function templates, and
you will need to implement them. You will also notice that there may be some changes or
modifications required to the data structures to support the necessary operations – feel free
to add member variables or functions, but please do not change the names of the provided
functions as these will be used for marking.
You will need to implement your own tests and run them using:
python3 test_structures.py --linkedlist.
Task 1.2: Dynamic Array (2 Marks)
Unlike the linked list discussed above, which can store nodes at any abitrary location in
memory, we often prefer to have data items stored contiguously (consecutively in memory),
allowing us to access an element x at some index i in constant time. One such way to achieve
this is through the use of a dynamic array.
The file dynamic_array.py contains another skeleton for you to implement. You should
store your data in self._data, and you can add any other member variables to your Dy-
namicArray object. Each function that needs to be supported is provided as a stub. Your
implementations should be efficient and correct, and we have provided annotations to describe
the expected complexity. In this assignment, we have given you a slightly trickier ADT
than the classic append-only array discussed in the lectures. In particular, you must support
prepend operations — that is, allowing an element x to be placed at the front of the array —
in O(1) amortized time, worst case.
Assignment 1 3
You will need to implement your own tests and run them using:
python3 test_structures.py --dynamicarray
Task 1.3: Bitvector (1.5 marks)
In some applications, it is useful to track the state of a collection of objects using simple
Boolean (True or False) flags. A na?ve way to do this is to simply use a (dynamic) array,
storing bool types as the underlying data. However, Boolean types are usually represented
by a machine word (** or 64 bits), meaning that we waste a lot of space with this approach.1
An alternative approach is to use a bitvector , which stores an array of b-bit integers to
represent each item — unset bits (value 0) represent False, and set bits (value 1) represent
True. Clearly, this approach uses 64× less space than the na?ve approach, as a single b = 64
bit integer can track the state of 64 items.
The file bit_vector.py contains another skeleton for you to implement. Note that it
uses a composition based design where a DynamicArray object is used to store the underlying
data. Each function that needs to be supported is provided as a stub. Your implementations
should be efficient and correct, and we have provided annotations to describe the expected
complexity. We describe two of the more exotic operations that should be supported in more
detail below.
You will need to implement your own tests and run them using:
python3 test_structures.py --bitvector.
Bitvector Operations: Shift
The shift operator handles both left and right shifts, depending on the sign of the dist
parameter. If the dist parameter is positive, we do a left shift by dist. A left shift moves all
bits in the bitvector left by dist positions, replacing empty positions with 0 bits. For example,
the following demonstrates the before (top) and after (bottom) of a left shift by dist = 2:
1011000100011
1100010001100
Notice that the two most significant bits have fallen off (the leftmost 10 on the first bitvector).
The right shifts work the same way, but we move the bits dist positions to the right (and the
least significant bits will fall off).
Bitvector Operations: Rotate
The rotate operator works exactly the same way as the shift operator, except it ensures that
any bits that fall off the end are rotated back onto the start of the bitvector. Using the same
example as above with dist = 2:
1011000100011
1100010001110
1 For the interested student: Most statically typed languages would represent a Boolean value as a
machine word (that is, ** or 64 bits) for convenience. Python is a dynamically typed language, and
each object actually carries a bunch of metadata around with it, and may be something like 24 bytes.
However, every True or False used is merely a reference to one of these two objects (there is only real
True and one real False object instantiated in the runtime). Nonetheless, here, we will pretend that
each Bool usually costs a machine word. In C, for example, a bool would use 8 bits (one byte), the
minimal addressable size.
COMP3506/7505
4 COMP3506/7505 – Semester 2, 2024
3 Task 2: Algorithmic Thinking Warm-Up (5 Marks)
Next, we are going to work on some simple warm up problems. These are designed to build
your problem solving skills. Some of them may appear tricky at first; you are encouraged
to sit down and think about them (a pen and a piece of paper will help). Do not be afraid
to get creative, as there may be multiple ways to solve each problem. Each problem will
be assessed on three tiers of tests — see the warmup.py file for more details, and test with
test_warmup.py (you need to implement your own tests).
3.1 The Main Character
You are given a string S. You need to simply return the first position of a repeated character
(indexed from zero), or ?1 if there are no repeats.
S = hello → 3.
S = world → ?1.
 S = algorithmsarefun → 10.
S = ooooohigetitnow → 1.
Here’s the catch: S can be built from an alphabet containing 2** possible characters,
represented as integers in the range [0, 2** ? 1]. Hint: You should use one of your data
structures from part one to help you solve this problem efficiently!
3.2 Sum-Thing Odd
You are given an unsorted list L containing n unique integers. Let min(L) and max(L)
represent the smallest and largest integers in L. Your task is to find the sum of the missing
odd integers in the range [min(L),max(L)].
 Consider L = ?6, 4, 9?: min(L) = 4 and max(L) = 9. Your range of interest is thus [4, 9].
The sum of the missing odd integers will be 5 + 7 = 12.
 Consider L = ?10, 1, 7, 17?: min(L) = 1 and max(L) = 17. Your range of interest is [1, 17].
The sum of missing odds in this range will be 3 + 5 + 9 + 11 + 13 + 15 = 56.
3.3 It’s cool, k?
A natural number is k-cool if it can be represented as the sum of unique non-negative powers
of k. For example:
? 17 is 4-cool because 40 + 42 = 17;
? 128 is 2-cool because 27 = 128;
? 11 is 10-cool because 100 + 101 = 11.
Given n and k, you must return the n th largest k-cool natural number. Two alternative
ways to say this:
? Return the n th smallest number that can be created by summing non-negative integer
powers of k; or
? If you created all combinations of sums of powers of k and sorted them into ascending
order, we want you to return the n th one.
Clearly, the first k-cool number is always 1 (k0 = 1) and the second k-cool number is always
k1 = k. Since we will test some very large numbers (up to 101,000,000), you should return the
number modulo 1016 + 61.
Assignment 1 5
3.4 Your Number is Up
Alice and Bob are playing a numbers game. The game starts with an unsorted list of integers,
L, with n = |L| guaranteed to be even. On each turn, they both remove one number from L.
If Alice removes an even number, it is added to her score. If she removes an odd number,
her score remains unchanged. The opposite is true for Bob (removing an odd number adds
to his score; removing an even number keeps his score unchanged). After n turns, the array
is exhausted, and the winner is determined by the player with the largest score — A draw is
also possible. Given L, you must return the winner, and their score, assuming both Alice
and Bob play optimally. You can assume that Alice always gets the first turn.
3.5 Getting Lit
A straight road of length k is illuminated by n light poles. Each light pole is represented
by a natural number i, where i represents the total distance from the start of the road to
the given pole. Each light is capable of illuminating a radius r. Given an unsorted list L of
poles, you need to return the minimal radius r such that the entire length of the road, k, is
illuminated. You may assume that the width of the road is infinitely small.2
k =12
L = 4, 7, 8, 1 .
0 3 6 9 12
r is too small!
0 3 6 9 12
r
r is too big!
0 3 6 9 12
r
r is just right.
0 3 6 9 12
r
Figure 2 A sketch of getting lit with n = 4 light poles.
2 Yes, we know it’s not a very useful road, but it makes your assignment easier. See also: https:
//w.wiki/Ajs5
COMP3506/7505
6 COMP3506/7505 – Semester 2, 2024
4 Task 3: Problem Solving with Data Structures (5 marks)
You are an algorithms specialist working at SIGSEGVTM, a world leader in high perform-
ance algorithmic solutions. You have been contracted by a bioinformatics company called
MallocLabs who require a bespoke system to help them deal with a growing amount of
genomic data they need to index. Their lead Bioinformatician, Barry Malloc, has provided
you with the following overview of the data and the system requirements. Your job is to
design and implement an appropriate data structure — and related algorithms — to match
Barry’s requirements. Barry has kindly placed the required time bounds in the function
stubs — you should carefully consider these when designing your data structure.
4.1 Data Representation
DNA data is represented as a string S of length |S| over an alphabet Σ = {A, C, G, T}.
Each character represents a different base (Adenine, Cytosine, Guanine, and Thymine). For
example, a sequence S with |S| = ** might look like
S = GTCGTGAAGTCGGTTCCTTCAATGGTTAAACC.
Since sequences can be very long, we can break them up into k-mers, all possible substrings
of length k. For example, there are 10 individual 23-mers of S:
GTCGTGAAGTCGGTTCCTTCAAT
TCGTGAAGTCGGTTCCTTCAATG
CGTGAAGTCGGTTCCTTCAATGG
GTGAAGTCGGTTCCTTCAATGGT
TGAAGTCGGTTCCTTCAATGGTT
GAAGTCGGTTCCTTCAATGGTTA
AAGTCGGTTCCTTCAATGGTTAA
AGTCGGTTCCTTCAATGGTTAAA
GTCGGTTCCTTCAATGGTTAAAC
TCGGTTCCTTCAATGGTTAAACC
In general, a sequence of length |S| will contain |S|?k+1 k-mers, and there are a total of |Σ|k
unique possible k-mers (in our case, |Σ| = 4). You are provided with a tool generate_dna.py
that can generate n sequences of length |S| for you to experiment with; it is probably easiest
to simply write them out to a file, and use the file as input to your testing program.
4.2 Required Functionality
At run-time, your program will be given two arguments that specify a path to a file containing
DNA sequences, as well as the value of k we are interested in working with. For example,
you might be given a file containing 50, 000 sequences of length 200, and k = 31. We will
always use str types to represent k-mers. The data structure used to solve the following
requirements is up to you, and should be designed based on the functionality requested. You
may need to use one or more of the structures implemented in part one for example, but the
final choice is yours. Implement in kmer_structure.py and test with test_kmers.py (you
need to implement your own tests).
Assignment 1 7
Storage and Modification: 2 marks
The first set of functions you need to support allow for reading and modifying data. They
are specified as follows:
? read: Given a file containing DNA sequences, break them into individual k-mers, and
store them in your data structure;
? batch_insert(L): Given a list of k-mers L, insert them into your data structure;
? batch_delete(L): Given a list of k-mers L, delete all occurences of them from your data
structure.
Note that there may be some duplicate k-mers in your data structure. You must keep track
of duplicates and their frequency, as these will be required for answering some query types in
the next section.
Queries: 3 marks
Your data structure also needs to support the following query types.
? freqgeq(n): Return a list of unique k-mers that occur at least n times;
? count(q): Return the number of times a k-mer q occurs;
? countgeq(q): Return the total number of k-mers that are ≥ q; that is, you need to sum
the frequencies of all k-mers lexicographically greater than or equal to q;
? compatible(q): Return the total number of k-mers that are compatible with q.
We provide some further information on the compatibility query as follows. A given k-mer q
is called compatible with k-mer b if the last two characters in q are the complement of the
first two characters in b. In genomics, the pair A and T is complementary, as is the pair C
and G. So, for example, CCTGATG is compatible with ACTTGCG:
q = CCTGATG
| |
ACTTGCG
Note that we always assume we are matching the end of the input query k-mer q with the
start of all other k-mers.
Analysis: 2 marks (COMP7505 Students Only)
If you are a COMP7505 student, you must also answer the questions posed in the plain text
file called analysis.txt (inside the malloclabs directory).COMP3506 students are encouraged
to do this too, but they will not be assessed on this component. Please keep your answers
succinct, but make sure to include all details that may be relevant. If in doubt, err on the
side of more detail.
COMP3506/7505
8 COMP3506/7505 – Semester 2, 2024
5 Assessment
This section briefly describes how your assignment will be assessed.
5.1 Mark Allocation
Marks will be provided based on an extensive (hidden) set of unit tests. These tests will do
their best to break your data structure in terms of time and/or correctness, so you need to
pay careful attention to the efficiency and the validity of your code. Each test passed will
carry some weight, and your autograder score will be computed based on the outcome of the
test suite. If you did not rigorously test your programs/code, you should go back and do so!
As the famous poet Ice Cube once said: check yourself before you wreck yourself.
The marks (percentages) provided in each task above are indicative of the total score
available for each part, but marks may be taken off for poor coding style including lack of
commenting, inefficient solutions, and incorrect solutions. Our code quality checks are not as
strict as PEP8, but we assume typical best practices are used such as informative variable
and function names, commenting, and breaking long lines. While the overall grade/score
will be calculated mathematically, an indicative rubric is provided as follows:
? Excellent: Passes at least **% of test cases, failing only sophisticated or tricky tests;
well structured and commented code; appropriate design choices; appropriate application
of data structures/algorithms for solving Tasks 2/3.
? Good: Passes at least 80% of test cases, failing one or two simple tests; well structured
and commented code; good design choices with some minor improvements possible;
good application of data structures/algorithms for solving Task 2/3 with some minor
improvements possible.
? Satisfactory: Passes at least 70% of test cases; code is reasonably well structured with
some comments; most design choices are reasonable but significant room for improvement;
reasonable application of data structures/algorithms for solving Task 2/3, but significant
improvements possible.
? Poor: Passes less than 70% of test cases; code is difficult to read, not well structured, or
lacks comments; design choices do not demonstrate a sound understanding of the desired
functionality; little or no suitable application of data structures or algorithms towards
solving Task 2/3.
5.2 Plagiarism and Generative AI
If you want to actually learn something in this course, our recommendation is that you avoid
using Generative AI tools: You need to think about what you are doing, and why, in order
to put the theory (what we talk about in the lectures and tutorials) into practical knowledge
that you can use, and this is often what makes things “click” when learning. Mindlessly
lifting code from an AI engine won’t teach you how to solve algorithms problems, and if
you’re not caught here, you’ll be caught soon enough by prospective employers.
If you are still tempted, note that we will be running your assignments through sophistic-
ated software similarity checking systems against a number of samples including including
your classmates and our own solutions (including a number that have been developed with
AI assistance). If we believe you may have used AI extensively in your solution, you may
be called in for an interview to walk through your code. Note also that the final exam
may contain questions or scenarios derived from those presented in the assignment work, so
cheating could weaken your chances of successfully passing the exam.
Assignment 1 9
As part of your submission, you must create a file called statement.txt. In that file,
you must provide attribution to any sources or tools used to help you with your assignment,
including any prompts provided to AI tooling. If you did not use any such tooling, you can
make a statement outlining that fact. Failing to submit this file will yield you zero marks.
6 Submission
You need to submit your solution to Gradescope under the Assignment 1: Autograder link in
your dashboard. Please use the appropriate link as there is a separate submission for 3506
and 7505 students. Once you submit your solution, a series of tests will be conducted and
the results of the public tests will be provided to you. However, the assessment will also
include a number of additional hidden tests, so you should make sure you test your solutions
extensively. You may resubmit as often as you like before the deadline, but we are imposing
a limit of ten submissions per 24 hour period. Please write your own tests!
Structure
The easiest way to submit your solution is to submit a .zip file. The autograder expects a
specific directory structure for your solution, and the tests will fail if you do not use this
structure. In particular, you should use the same structure as the skeleton codebase that
was provided. You should also have the statement.txt and analysis.txt (for COMP7505
students). Submissions without the statement.txt will be given zero marks, and the
autograder will notify you of this.
7 Resources
We provide a number of useful git and/or unix resources to help you get started. Please go
onto the Blackboard LMS and see the Learning Resources > Resources directory for more
information.
8 Changelog
? V1.0: Initial release.
? V2.0: Release with code; changed directory structure to match code. Clarified k-cool
further. Fixed an unfortunate typo.
? V3.0: Add clarifications section to the end of the spec to track any additonal changes or
clarifications from Ed or other discussions.
? V3.1: Remove some of the clarifying discussion about linked list nodes, since the API
was changed in the v3.0 code skeleton and the discussion is no longer relevant.
? V4.0: Further clarifications in the section below. No further changes.
COMP3506/7505
10 COMP3506/7505 – Semester 2, 2024
9 Clarifications
This section is introduced in V3.0 of the spec. It will be used to track any additional
clarifications on the spec or functionality.
Corner Cases
Here we clarify some corner cases and expected behaviour.
DoublyLinkedList
? What happens if I call set_head (or set_tail) on an empty list? In this case, do nothing.
These functions simply change the data of the head or tail if they do already exist. The
v3 skeleton code docstrings now capture this functionality.
BitVector
? If the dist provided to shift is greater than the bitvector length, what happens? The
bitvector would simply become filled with 0’s as everything will be shifted off.
? What about the case where we call rotate with a dist value larger than the length?
You will just keep rotating until done; rotating a bitvector of length l by l or 2l or 3l will
end up with the same bitvector as before the rotation. Modulus is your friend...
Node API
In our test suite, we will not be checking individual linked list Node types in isolation. What we
will be checking is that the linked list API works as expected; we will be traversing/modifying
your linked list (via the public DoublyLinkedList methods) and comparing the behaviour with
what a correct implementation does.
So, to be clear, we will never take a single Node type and check what is in prev or next
for example. This means you may even modify the Node API as it is only accessed internally
from the DoublyLinkedList; just ensure it is still compatible with the DoublyLinkedList
API.
Banned Code
Our philosophy for banning specific python functionality is to avoid the situation where some
complex operations are being hidden by syntactic sugar. For example, consider:
my_list = [1, 6, 105, 4, 9]
x = 4
if x in my_list:
print ("Yay!")
This innocent looking code is actually hiding an O(n) execution, where n is the length of
my_list. That’s because Python allows us to search for an item using nice syntax like if x
in my_list. That’s why we would prefer you to write something like:
...
for item in my_list:
if x == item:
print ("Yay!")
Even though this is longer, it clearly demonstrates that the list is being iterated over.
Some examples of things that are OK or not:
Assignment 1 11
? Generators: These are OK, because they hide annoying/ugly complexity, but do not mask
the fact that they are used for iteration.
? Dictionaries: Clearly not OK, because you need to be able to implement a map data
structure to get this behaviour. Same for Sets. (More in Week 5/6/7).
? Dunder methods: These are OK, as you will need to implement them anyway (and indeed,
you will implement __str__ for example).
In general: If you want to use something, you need to implement it yourself, not use an “out
of the box” implementation.
Why Moss?
You do not need to use Moss. We recommend you do because it provides a shared platform
for us to help if anything goes wrong. However, you are more than welcome to develop/test
your work elsewhere.
More about Bitvectors
A key principle in this assignment is that the data structure being implemented should
behave the way a user expects. This is different to the question of how the data is stored (in
what sort of container, in what order, ...) as a user doesn’t care about that. We do care, as
the designer of the data structure, so we endeavour to make things as efficient as possible
while maintaining correctness.
In terms of bitvectors, we cannot answer questions like should I append a bit to the most
significant bit, or the least significant bit? or should I store my bits little or big endian? –
What really matters is that a user appending the sequence 1 0 1 0 will get back a 1 if they
ask for the value of the bit at index 0, and a 0 back if they ask for the value of the bit at
index 3. If they then flip all of the bits, and prepend a 0, they should get a 0 back if they ask
for the value at index 0, and a 0 if they ask for the element at index 1 (this was previously
the first 1 we appended, but it has now been flipped). When these operations happen, it
is up to you to design how it works under the hood. The API should behave like a user
interacting with the DynamicArray using only 0 and 1’s, except that it should be much more
space efficient.
I’m Overwhelmed
This is a tough assignment, and it will take time. If you are overwhelmed, our advice is to
simplify the API you support for each structure. For example, instead of supporting both
append and prepend, just focus on supporting append efficiently. Similarly, ignore reverse for
now. This will at least allow you to build an efficient “append-only” list that you can use in
the assignment. You will of course lose some marks, but you will be able to move on and
capitalise in the other sections. For the warmup problems, just focus on solving them – even
if it is inefficient at first – as you will get marks just for giving correct (albeit slow) solutions
on small inputs.
Finally, please seek support; we have Ed, office hours, assignment help sessions. But we
cannot help you if you do not seek help.
COMP3506/7505

請加QQ:99515681  郵箱:99515681@qq.com   WX:codinghelp







 

掃一掃在手機打開當(dāng)前頁
  • 上一篇:代寫 ACCT90004、代做 C++,Python 編程語言
  • 下一篇:COMP90025 代做、代寫 C/C++編程語言
  • 無相關(guān)信息
    合肥生活資訊

    合肥圖文信息
    急尋熱仿真分析?代做熱仿真服務(wù)+熱設(shè)計優(yōu)化
    急尋熱仿真分析?代做熱仿真服務(wù)+熱設(shè)計優(yōu)化
    出評 開團(tuán)工具
    出評 開團(tuán)工具
    挖掘機濾芯提升發(fā)動機性能
    挖掘機濾芯提升發(fā)動機性能
    海信羅馬假日洗衣機亮相AWE  復(fù)古美學(xué)與現(xiàn)代科技完美結(jié)合
    海信羅馬假日洗衣機亮相AWE 復(fù)古美學(xué)與現(xiàn)代
    合肥機場巴士4號線
    合肥機場巴士4號線
    合肥機場巴士3號線
    合肥機場巴士3號線
    合肥機場巴士2號線
    合肥機場巴士2號線
    合肥機場巴士1號線
    合肥機場巴士1號線
  • 短信驗證碼 豆包 幣安下載 AI生圖 目錄網(wǎng)

    關(guān)于我們 | 打賞支持 | 廣告服務(wù) | 聯(lián)系我們 | 網(wǎng)站地圖 | 免責(zé)聲明 | 幫助中心 | 友情鏈接 |

    Copyright © 2025 hfw.cc Inc. All Rights Reserved. 合肥網(wǎng) 版權(quán)所有
    ICP備06013414號-3 公安備 42010502001045

    99爱在线视频这里只有精品_窝窝午夜看片成人精品_日韩精品久久久毛片一区二区_亚洲一区二区久久

          9000px;">

                亚洲午夜久久久久| 五月激情综合色| 亚洲电影中文字幕在线观看| 色诱亚洲精品久久久久久| 亚洲愉拍自拍另类高清精品| 欧美一区二区免费观在线| 国产乱码精品一区二区三区av| 久久综合久色欧美综合狠狠| 不卡区在线中文字幕| 亚洲女与黑人做爰| 欧美高清激情brazzers| 国产一区视频在线看| 亚洲欧美成人一区二区三区| 日韩三级视频中文字幕| youjizz久久| 免费美女久久99| 亚洲三级在线免费| 久久亚洲免费视频| 在线精品视频小说1| 久久国产精品99久久人人澡| 亚洲欧美日韩久久精品| 精品久久久久av影院| 91久久精品一区二区三| 国产在线精品一区二区不卡了| 亚洲靠逼com| 久久久久久免费| 日韩欧美在线1卡| 91精品91久久久中77777| 国产精品一品二品| 青娱乐精品视频| 亚洲国产成人av| 亚洲少妇屁股交4| 国产香蕉久久精品综合网| 7777精品伊人久久久大香线蕉的| 99久久精品国产毛片| 国产精品1024| 精品一区二区三区不卡| 天堂av在线一区| 午夜激情一区二区三区| 一区二区三区在线免费播放| 国产精品乱人伦中文| 国产欧美日产一区| 久久伊人中文字幕| 欧美α欧美αv大片| 日本不卡视频在线| 国产91综合一区在线观看| 亚洲午夜久久久久久久久久久| 欧美激情一区在线观看| 欧美精品一区二| 久久综合狠狠综合| 精品国产百合女同互慰| 精品免费视频.| 久久久噜噜噜久久中文字幕色伊伊| 欧美大片一区二区| 亚洲精品在线一区二区| 欧美精品一区二区三区蜜臀| 日韩视频免费观看高清完整版| 日韩视频在线你懂得| 色一情一伦一子一伦一区| 久久精品久久久精品美女| 久久99精品国产.久久久久| 精品在线一区二区| 国产盗摄精品一区二区三区在线| 国产福利精品导航| 成人午夜看片网址| 色网站国产精品| 欧美精品三级在线观看| 精品久久久久久综合日本欧美| 欧美不卡一区二区三区| 欧美国产一区在线| 亚洲美女视频在线| 午夜精品一区二区三区电影天堂| 日本视频中文字幕一区二区三区| 久久成人免费网| 成人av影院在线| 欧美性猛交xxxxxxxx| 欧美一区在线视频| 国产精品色噜噜| 午夜精品一区二区三区电影天堂| 美女国产一区二区三区| 成人激情午夜影院| 91精品欧美久久久久久动漫| 中文字幕+乱码+中文字幕一区| 亚洲国产中文字幕在线视频综合 | 97成人超碰视| 欧美丰满少妇xxxbbb| 国产亚洲一区二区三区| 亚洲国产美国国产综合一区二区 | 免费看精品久久片| 成人国产亚洲欧美成人综合网| 欧美日韩亚洲综合在线| 欧美激情一区二区三区在线| 亚洲一区二区av在线| 国产成人在线影院| 欧美成人aa大片| 亚洲一区二区三区四区不卡| 国产91在线|亚洲| 欧美本精品男人aⅴ天堂| 亚洲精品免费播放| 粉嫩欧美一区二区三区高清影视| 欧美久久久久久久久| 亚洲天堂av一区| 成人午夜免费视频| 久久精品视频在线看| 人禽交欧美网站| 欧美性视频一区二区三区| 国产精品天天摸av网| 久久国产精品72免费观看| 欧美日韩免费不卡视频一区二区三区| 国产精品午夜久久| 国产成人综合亚洲91猫咪| 欧美不卡一二三| 青椒成人免费视频| 51午夜精品国产| 婷婷综合另类小说色区| 在线视频亚洲一区| **欧美大码日韩| 成人在线一区二区三区| 欧美色综合影院| 国产精品你懂的在线| 奇米影视一区二区三区| 在线免费观看日本一区| 国产精品国模大尺度视频| 久久精品国产色蜜蜜麻豆| 欧美精品1区2区| 视频一区在线视频| 欧美日韩国产美女| 五月激情综合色| 欧美精品精品一区| 午夜精品一区在线观看| 欧美在线你懂得| 日韩国产欧美在线观看| 日韩欧美成人午夜| 国产精品一区不卡| 中文字幕欧美一区| 色综合久久久久久久| 亚洲色图自拍偷拍美腿丝袜制服诱惑麻豆| 国产91精品久久久久久久网曝门| 欧美亚洲图片小说| 久久久久久免费| 国产91精品入口| 国产精品久久免费看| 91麻豆自制传媒国产之光| 亚洲色欲色欲www| 欧美图片一区二区三区| 五月天激情小说综合| 日韩情涩欧美日韩视频| 精品一区二区成人精品| 国产三级三级三级精品8ⅰ区| 国产成人综合自拍| 亚洲精品福利视频网站| 欧美久久久久久久久中文字幕| 蜜臀久久久久久久| 久久日韩精品一区二区五区| 99久久精品国产导航| 一区二区三区精品视频| 欧美一区二区三区日韩视频| 久久av老司机精品网站导航| 中日韩av电影| 9191久久久久久久久久久| 成人免费观看男女羞羞视频| 夜夜亚洲天天久久| 欧美日韩不卡视频| 奇米精品一区二区三区在线观看 | 国产情人综合久久777777| 99久久精品免费精品国产| 午夜精品福利久久久| 欧美激情在线一区二区| 欧美人与z0zoxxxx视频| 粉嫩欧美一区二区三区高清影视| 亚洲国产欧美另类丝袜| 久久久久国产精品麻豆| 欧美日韩一区二区欧美激情| 高潮精品一区videoshd| 爽好久久久欧美精品| 亚洲青青青在线视频| 久久久91精品国产一区二区精品 | 日韩美女视频在线| 91美女精品福利| 激情综合五月天| 日韩电影一区二区三区四区| 一区二区三区在线免费| 国产精品免费免费| 精品成人在线观看| 91精品国产一区二区三区香蕉| 99久久婷婷国产精品综合| 韩国精品在线观看| 免费看日韩精品| 日韩经典中文字幕一区| 亚洲成人av在线电影| 一区二区三区中文字幕| 欧美伊人久久大香线蕉综合69 | 国模套图日韩精品一区二区 | 欧美日韩激情一区二区三区| 成人免费视频网站在线观看| 国内不卡的二区三区中文字幕 | 日韩av一区二区在线影视| 亚洲乱码一区二区三区在线观看| 国产精品妹子av| 中文字幕一区二区三区乱码在线|