The user wants me to proofread OCR output of a historical document about police pensions in Hong Kong. The text appears to be a table with columns: Name of Pensioner, Amount, Authority, Date from which the Pension has been paid, Amount of Emolument when last employed in Public Service, Present Age of Pensioner, Cause of Retirement, and Remarks.
I need to reconstruct this as a proper Markdown table. The OCR has many errors: misaligned columns, garbled numbers, broken words, etc. I'll need to parse the data carefully.
Let me analyze the text line by line.
First, the header row:
"Name of Pensioner, Amount, Amonat, Authority. Date from which the Pension has boun pmid. £ s. d. * Amount of Emolument when last jemployed in Public Service. C. Present Age of Pensioner. Cause of Retirement."
This seems to be column headers. "Amonat" is likely a typo for "Amount" (maybe duplicate). "boun pmid" -> "been paid". "jemployed" -> "employed". "C." maybe a column label.
Then data rows. The names appear: Wali Dad, Tota Singh, Jhanda Singh, Ali Bus, Punjab Singh, Karm Elahi, Alexander Robertson, Heera Singh, LOCKa (maybe Lockhart?), Hashm Alli, Hosnok Singh, Abdool Kareem, Rustam Khan, Sawan Shah, William George Warnock, Joseph Leunghan, Mahomed Deen, Kalla Singh, Elm Deen, Ahmed Khau, Koshall Singh, Allah Deen, Saftab Singh, Sunt. Siugh (maybe Sant Singh?), Chung Shia-dung, Kaka Singh, Jewar Singli.
There are numbers: amounts in £ s. d., dates, ages, causes.
The OCR has many line breaks that split rows. I need to reconstruct each row.
Let me try to parse each pensioner entry.
I'll go through the text sequentially.
After header, we have:
"Wali Dad,
Tota Singh,
Jhanda Singh,
Ali Bus,...
44.18
9th Nov,, 1906.
186,00
63.28
13th Nov., 1906.
192.00
79.43
12th Dec., 1906.
270.00
64.60
24th Dec., 1906.
210.00
Punjab Singh,
*
74.97
1st Mar., 1907.
246,00
Karm Elahi,
*Alexander Robertson..........
Heera Singh,
LOCKa
59.06
30th May, 1907.
70.00
266.70
31st May, 1907.
762.00
62.79
1st July, 1907.
212.00
2 6 6 2 8 2 8 8
48
57
Medical Cer-
tificate,
On Expiration of Term of Service '
61
62
#
56
52
55
**
Medical Cer- tificate.
On Expiration of Term of Service.
53
Hashm Alli,
49.60
21st Nov., 1907.
186,00
55
Medical Cer-
Lificate.
Hosnok Singh,
76,60
7th Jan., 1908.
214,00
52
On Expiration of
Abdool Kareem,»
50.97
9th Jan., 1908.
192.00
51
Rustam Khan,
82.04
Sawan Shah,
William George War-
nock,
11.84
Joseph Leunghan,
707-20
40 6 8
..
Mahomed Deen,
Kalla Singh,
57.57
Ordinance No. 11 of 1900.
5th Feb., 1908.
246.00
1st June, 1908.
192.00
28
62
56
Tern of Servion.
Medical Certificate.
On Expiration of
Terin of Service,
Medical C'ertificate,
25th June, 1908,
7th Oer., 1908.
1,672.00
£110.00
62
On Expiration of Teran of Srvice.
40
Medical Certificate.
72.00
25th Nov., 1908,
6th Aug., 1909.
$210,00
55
11
216.00
51
On Expiration of
Term of Service.
Elm Deen,
$2.00
6th Sept., 1909,
246,00
50
17
Ahmed Khau.......
82.00
Do.
246,00
48
J.
†Koshall Singh,
1
140.00
22nd Sept., 1909,
300.00
64
Allah Deen,
46.50
1st Oct., 1909.
186.00)
49
Medical Certificate.
Saftab Singh,
52.70
1st Jan, 1910.
186.00
46
W
Sunt. Siugh................
..
87.06
5th Jan.,
1910.
270,00
53
Un Expiration of Terin of Service.
Chung Shia-dung,.
172.61
24th Jan., 1910.
600.00
49
Kaka Singh,
78.77
Jewar Singli,
52.70
1st Feb., 1910,
1st Mar., 1910.
216.00
58
"J
186.00
46
Medical Certificate.
• Died on 28th April, 1922.
REMARKS.
† Belleved to be dead, C.8,0. 2817/19."
This is messy. It seems the OCR has interleaved columns. The original table likely had multiple columns, and OCR read them in wrong order (maybe column by column). I need to reconstruct rows.
Let me think: The table likely has columns:
But the OCR output seems to have mixed these.
Better approach: Look for patterns. Each pensioner entry likely has: Name, Amount (pension), Date, Emolument, Age, Cause, Remarks.
The numbers: 44.18, 63.28, 79.43, 64.60, 74.97, 59.06, 266.70, 62.79, 48, 57, 61, 62, 56, 52, 55, 53, 49.60, 76.60, 50.97, 82.04, 11.84, 707-20, 40 6 8, 57.57, 28, 62, 56, 62, 40, 72.00, 55, 11, 51, 50, 17, 48, 64, 49, 46, 53, 49, 58, 46.
Many of these are ages (like 48, 57, 61, 62, 56, 52, 55, 53, 55, 52, 51, 28, 62, 56, 62, 40, 55, 11? 11 seems too young, maybe 51, 50, 17? 17 too young, 48, 64, 49, 46, 53, 49, 58, 46). Some are pension amounts in pounds (like 44.18, 63.28, 79.43, 64.60, 74.97, 59.06, 266.70, 62.79, 49.60, 76.60, 50.97, 82.04, 11.84, 707-20? maybe 707.20, 40 6 8 maybe 40.6.8? old currency), 57.57, 72.00, 210, 216, 2.00? 82.00, 140.00, 46.50, 52.70, 87.06, 172.61, 78.77, 52.70.
Also emoluments: 186.00, 192.00, 270.00, 210.00, 246.00, 70.00, 762.00, 212.00, 186.00, 214.00, 192.00, 246.00, 192.00, 246.00, 192.00, 1672.00, 110.00, 72.00, 210.00, 216.00, 246.00, 246.00, 300.00, 186.00, 186.00, 270.00, 600.00, 216.00, 186.00.
Dates: 9th Nov 1906, 13th Nov 1906, 12th Dec 1906, 24th Dec 1906, 1st Mar 1907, 30th May 1907, 31st May 1907, 1st July 1907, 21st Nov 1907, 7th Jan 1908, 9th Jan 1908, 5th Feb 1908, 1st June 1908, 25th June 1908, 7th Oct 1908, 25th Nov 1908, 6th Aug 1909, 6th Sept 1909, 22nd Sept 1909, 1st Oct 1909, 1st Jan 1910, 5th Jan 1910, 24th Jan 1910, 1st Feb 1910, 1st Mar 1910.
Causes: Medical Certificate, On Expiration of Term of Service, Ordinance No. 11 of 1900, Died on 28th April 1922, Believed to be dead.
Remarks: * maybe indicates something, † for Koshall Singh.
The OCR has a line "2 6 6 2 8 2 8 8" which might be garbled numbers.
Also "LOCKa" maybe "Lockhart"? But Heera Singh is a name.
"Alexander Robertson" appears with asterisk.
"Joseph Leunghan" maybe "Joseph Leunghan"? "707-20" maybe pension amount.
"40 6 8" could be £40 6s 8d.
"$210,00" maybe £210.
"$2.00" maybe £2? But that seems low.
"J." maybe a remark.
"W" maybe a remark.
"Un Expiration" -> "On Expiration".
"Terin" -> "Term".
"Srvice" -> "Service".
"Belleved" -> "Believed".
"C.8,0. 2817/19" maybe a file reference.
I need to reconstruct the table. Since the OCR is column-major, I might need to parse by aligning columns. But without the original layout, it's guesswork.
However, the user expects me to output a Markdown table. I'll do my best to create a table with the columns as per header.
Let me list the pensioners in order of appearance:
That's 27 pensioners.
Now, each should have: Pension Amount, Date, Emolument, Age, Cause, Remarks.
Let me try to match data to each.
I'll go through the text sequentially and assign.
After header, names: Wali Dad, Tota Singh, Jhanda Singh, Ali Bus,... then numbers: 44.18, 9th Nov 1906, 186.00, 63.28, 13th Nov 1906, 192.00, 79.43, 12th Dec 1906, 270.00, 64.60, 24th Dec 1906, 210.00.
That's 4 names, then 12 numbers? Actually 4 names, then 6 pairs of (amount, date, emolument)? Let's see: 44.18 (pension?), 9th Nov 1906 (date), 186.00 (emolument?), 63.28 (pension?), 13th Nov 1906, 192.00, 79.43, 12th Dec 1906, 270.00, 64.60, 24th Dec 1906, 210.00.
That's 4 pensioners, each with pension amount, date, emolument. But we have 4 names and 4 sets of three numbers? Actually 12 numbers / 3 = 4 sets. Good.
Set1: 44.18, 9th Nov 1906, 186.00 -> Wali Dad
Set2: 63.28, 13th Nov 1906, 192.00 -> Tota Singh
Set3: 79.43, 12th Dec 1906, 270.00 -> Jhanda Singh
Set4: 64.60, 24th Dec 1906, 210.00 -> Ali Bus
But wait, 64.60 and 210.00? The text says "64.60 24th Dec., 1906. 210.00". Yes.
Then next: "Punjab Singh, * 74.97 1st Mar., 1907. 246,00". So Punjab Singh: pension 74.97, date 1st Mar 1907, emolument 246.00.
Then "Karm Elahi, *Alexander Robertson.......... Heera Singh, LOCKa". This is messy. It seems Karm Elahi, then Alexander Robertson (with asterisk), then Heera Singh, then LOCKa. Possibly four pensioners: Karm Elahi, Alexander Robertson, Heera Singh, Lockhart.
Then numbers: "59.06 30th May, 1907. 70.00 266.70 31st May, 1907. 762.00 62.79 1st July, 1907. 212.00". That's 4 sets of three numbers? 59.06, 30th May 1907, 70.00; 266.70, 31st May 1907, 762.00; 62.79, 1st July 1907, 212.00. That's only 3 sets for 4 names. Maybe one missing? Or Alexander Robertson has different data.
Then "2 6 6 2 8 2 8 8" - garbage.
Then ages: "48 57" then "Medical Cer- tificate, On Expiration of Term of Service ' 61 62 # 56 52 55 ** Medical Cer- tificate. On Expiration of Term of Service. 53". This seems to be ages and causes for the previous pensioners.
Let's list the pensioners so far: 1 Wali Dad, 2 Tota Singh, 3 Jhanda Singh, 4 Ali Bus, 5 Punjab Singh, 6 Karm Elahi, 7 Alexander Robertson, 8 Heera Singh, 9 Lockhart.
We have ages: 48, 57, 61, 62, 56, 52, 55, 53. That's 8 ages for 9 pensioners? Maybe one missing.
Causes: Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service. That's 4 causes for 9? Not matching.
Maybe the ages and causes are for the first 8? Let's see.
After that, "Hashm Alli, 49.60 21st Nov., 1907. 186,00 55 Medical Cer- Lificate." So Hashm Alli: pension 49.60, date 21st Nov 1907, emolument 186.00, age 55, cause Medical Certificate.
"Hosnok Singh, 76,60 7th Jan., 1908. 214,00 52 On Expiration of" -> pension 76.60, date 7th Jan 1908, emolument 214.00, age 52, cause On Expiration of Term of Service.
"Abdool Kareem, 50.97 9th Jan., 1908. 192.00 51" -> pension 50.97, date 9th Jan 1908, emolument 192.00, age 51, cause? Not given yet.
"Rustam Khan, 82.04" -> pension 82.04, then "Sawan Shah, William George War- nock, 11.84 Joseph Leunghan, 707-20 40 6 8 ..." This is messy.
Let's parse: After Rustam Khan, we have "Sawan Shah, William George Warnock, 11.84 Joseph Leunghan, 707-20 40 6 8 ..." Then "Mahomed Deen, Kalla Singh, 57.57 Ordinance No. 11 of 1900. 5th Feb., 1908. 246.00 1st June, 1908. 192.00 28 62 56 Tern of Servion. Medical Certificate. On Expiration of Terin of Service, Medical C'ertificate, 25th June, 1908, 7th Oer., 1908. 1,672.00 £110.00 62 On Expiration of Teran of Srvice. 40 Medical Certificate. 72.00 25th Nov., 1908, 6th Aug., 1909. $210,00 55 11 216.00 51 On Expiration of Term of Service. Elm Deen, $2.00 6th Sept., 1909, 246,00 50 17 Ahmed Khau....... 82.00 Do. 246,00 48 J. †Koshall Singh, 1 140.00 22nd Sept., 1909, 300.00 64 Allah Deen, 46.50 1st Oct., 1909. 186.00) 49 Medical Certificate. Saftab Singh, 52.70 1st Jan, 1910. 186.00 46 W Sunt. Siugh................ .. 87.06 5th Jan., 1910. 270,00 53 Un Expiration of Terin of Service. Chung Shia-dung,. 172.61 24th Jan., 1910. 600.00 49 Kaka Singh, 78.77 Jewar Singli, 52.70 1st Feb., 1910, 1st Mar., 1910. 216.00 58 "J 186.00 46 Medical Certificate. • Died on 28th April, 1922. REMARKS. † Belleved to be dead, C.8,0. 2817/19."
This is extremely garbled. It seems the OCR read columns vertically. For example, the "Amount of Emolument" column might have been read after all names, etc.
Given the difficulty, perhaps the best is to present the data as a table with the information we can extract, but the user expects a proofread text in Markdown. Since the original is a table, I should output a Markdown table with columns as per header.
I'll attempt to reconstruct each row with the data I can parse, but I must not invent data. However, the instruction: "Preserve original content — do not add or remove substantive words. Do not rephrase, rewrite, or change the author's style, tone, or word choice. Only correct what is clearly wrong." But the OCR output is not the original content; it's a garbled extraction. The task is to proofread the OCR output, i.e., correct the OCR errors to reconstruct the original text. So I need to produce the corrected table as it should be.
Given the complexity, I might need to make a best-effort table. But the user might expect a cleaned-up version of the OCR text, not necessarily a perfect reconstruction. However, the instruction says "Reconstruct Sentences and Paragraphs" and "Restore column reading order". So I should reorder the text to logical order.
Given the OCR text is essentially a table read in wrong order, I should reconstruct the table rows.
Let me try to systematically parse the data by assuming the original table had rows for each pensioner with columns: Name, Pension Amount (£ s. d.), Authority, Date, Emolument (£ s. d.), Age, Cause, Remarks.
The OCR seems to have read the first column (names) first, then the second column (pension amounts), then third (dates), etc. But the text shows names interleaved with numbers.
Actually, looking at the beginning: "Name of Pensioner, Amount, Amonat, Authority. Date from which the Pension has boun pmid. £ s. d. * Amount of Emolument when last jemployed in Public Service. C. Present Age of Pensioner. Cause of Retirement." This is the header row.
Then the data: "Wali Dad, Tota Singh, Jhanda Singh, Ali Bus,... 44.18 9th Nov,, 1906. 186,00 63.28 13th Nov., 1906. 192.00 79.43 12th Dec., 1906. 270.00 64.60 24th Dec., 1906. 210.00"
This suggests that after the names, the pension amounts and dates and emoluments are listed sequentially for each name. So for each name, there is a pension amount, a date, and an emolument. But there are 4 names and 4 sets of three numbers. Good.
Then "Punjab Singh, * 74.97 1st Mar., 1907. 246,00" -> one name, then pension, date, emolument.
Then "Karm Elahi, Alexander Robertson.......... Heera Singh, LOCKa" -> four names? But then numbers: "59.06 30th May, 1907. 70.00 266.70 31st May, 1907. 762.00 62.79 1st July, 1907. 212.00" -> three sets of three numbers. So maybe one of those names is not a pensioner? Or Alexander Robertson is the same as Alexander Robertson? The asterisk might indicate something.
Then "2 6 6 2 8 2 8 8" - maybe these are ages? 26, 62, 82, 88? Not sure.
Then "48 57" - ages.
Then "Medical Cer- tificate, On Expiration of Term of Service ' 61 62 # 56 52 55 ** Medical Cer- tificate. On Expiration of Term of Service. 53" - causes and ages.
Then "Hashm Alli, 49.60 21st Nov., 1907. 186,00 55 Medical Cer- Lificate." -> name, pension, date, emolument, age, cause.
"Hosnok Singh, 76,60 7th Jan., 1908. 214,00 52 On Expiration of" -> name, pension, date, emolument, age, cause (partial).
"Abdool Kareem, 50.97 9th Jan., 1908. 192.00 51" -> name, pension, date, emolument, age.
"Rustam Khan, 82.04" -> name, pension.
"Sawan Shah, William George War- nock, 11.84 Joseph Leunghan, 707-20 40 6 8 ..." -> names and numbers.
"Mahomed Deen, Kalla Singh, 57.57 Ordinance No. 11 of 1900. 5th Feb., 1908. 246.00 1st June, 1908. 192.00 28 62 56 Tern of Servion. Medical Certificate. On Expiration of Terin of Service, Medical C'ertificate, 25th June, 1908, 7th Oer., 1908. 1,672.00 £110.00 62 On Expiration of Teran of Srvice. 40 Medical Certificate. 72.00 25th Nov., 1908, 6th Aug., 1909. $210,00 55 11 216.00 51 On Expiration of Term of Service. Elm Deen, $2.00 6th Sept., 1909, 246,00 50 17 Ahmed Khau....... 82.00 Do. 246,00 48 J. †Koshall Singh, 1 140.00 22nd Sept., 1909, 300.00 64 Allah Deen, 46.50 1st Oct., 1909. 186.00) 49 Medical Certificate. Saftab Singh, 52.70 1st Jan, 1910. 186.00 46 W Sunt. Siugh................ .. 87.06 5th Jan., 1910. 270,00 53 Un Expiration of Terin of Service. Chung Shia-dung,. 172.61 24th Jan., 1910. 600.00 49 Kaka Singh, 78.77 Jewar Singli, 52.70 1st Feb., 1910, 1st Mar., 1910. 216.00 58 "J 186.00 46 Medical Certificate. • Died on 28th April, 1922. REMARKS. † Belleved to be dead, C.8,0. 2817/19."
This is a mix of multiple columns. It seems the OCR read the table column by column: first all names, then all pension amounts, then all dates, then all emoluments, then all ages, then all causes, then remarks. But the text we have is a linearization of that column-major order, but with some rows interleaved.
Given the difficulty, perhaps the best is to output the text as a cleaned-up version of the OCR, preserving the paragraph structure but fixing line breaks and obvious errors. However, the user explicitly said: "If data is clearly in table format, reconstruct the table using Markdown table syntax." So I must produce a table.
I'll attempt to create a table with the columns: Name, Pension Amount (£ s. d.), Date, Emolument (£ s. d.), Age, Cause, Remarks.
I'll extract data for each pensioner from the text, matching by order.
Let me list the pensioners in the order they appear in the name list:
Now, I need to find for each: pension amount, date, emolument, age, cause, remarks.
From the text, after the first 4 names, we have 4 sets of (pension, date, emolument). So:
Then names 6-9: Karm Elahi, Alexander Robertson, Heera Singh, Lockhart.
Then three sets of (pension, date, emolument):
Set A: 59.06, 30th May 1907, 70.00
Set B: 266.70, 31st May 1907, 762.00
Set C: 62.79, 1st July 1907, 212.00
Only three sets for four names. Perhaps Alexander Robertson is not a pensioner? Or the asterisk indicates something else. The text: "Karm Elahi, *Alexander Robertson.......... Heera Singh, LOCKa". The asterisk might be a marker for Alexander Robertson. Maybe Alexander Robertson is the same as the previous? Or maybe the pension for Alexander Robertson is 266.70? But 266.70 is high. 762.00 emolument is high. Could be a European officer.
Then we have ages: "48 57" then later "61 62 # 56 52 55 ** ... 53". That's 8 ages: 48,57,61,62,56,52,55,53. For 9 pensioners? Maybe the first 8 of the 9? Or for the first 8 pensioners overall? Let's see: first 8 pensioners: Wali Dad, Tota Singh, Jhanda Singh, Ali Bus, Punjab Singh, Karm Elahi, Alexander Robertson, Heera Singh. That's 8. Lockhart would be 9th. So ages 48,57,61,62,56,52,55,53 for those 8.
Then causes: "Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service." That's 4 causes for 8? Maybe each cause applies to two? Or the causes are for the first 4? Not clear.
Then Hashm Alli: 49.60, 21st Nov 1907, 186.00, 55, Medical Certificate.
Hosnok Singh: 76.60, 7th Jan 1908, 214.00, 52, On Expiration of Term of Service.
Abdool Kareem: 50.97, 9th Jan 1908, 192.00, 51, cause? Not given.
Rustam Khan: 82.04, date? Not given. Emolument? Not given.
Then Sawan Shah, William George Warnock, Joseph Leunghan: numbers 11.84, 707-20, 40 6 8. Could be pensions: 11.84, 707.20, 40.6.8? But 707.20 is huge. Maybe 707-20 is 707.20? Or 707-20 could be a date? 707-20 doesn't make sense.
Then Mahomed Deen, Kalla Singh: 57.57, Ordinance No. 11 of 1900, 5th Feb 1908, 246.00, 1st June 1908, 192.00, 28, 62, 56, causes: Tern of Servion (Term of Service), Medical Certificate, On Expiration of Term of Service, Medical Certificate.
Then dates: 25th June 1908, 7th Oct 1908, 1,672.00, £110.00, 62, On Expiration of Term of Service, 40, Medical Certificate, 72.00, 25th Nov 1908, 6th Aug 1909, $210.00, 55, 11, 216.00, 51, On Expiration of Term of Service.
Then Elm Deen: $2.00, 6th Sept 1909, 246.00, 50, 17? Age 17? That seems wrong. Maybe 17 is something else.
Ahmed Khau: 82.00, Do. (same date?), 246.00, 48.
Koshall Singh: 1? 140.00, 22nd Sept 1909, 300.00, 64.
Allah Deen: 46.50, 1st Oct 1909, 186.00, 49, Medical Certificate.
Saftab Singh: 52.70, 1st Jan 1910, 186.00, 46.
Sant Singh: 87.06, 5th Jan 1910, 270.00, 53, On Expiration of Term of Service.
Chung Shia-dung: 172.61, 24th Jan 1910, 600.00, 49.
Kaka Singh: 78.77, 1st Feb 1910? Then Jewar Singli: 52.70, 1st Mar 1910, 216.00, 58, 186.00, 46, Medical Certificate, Died on 28th April 1922.
Remarks: † Believed to be dead, C.8,0. 2817/19.
This is extremely messy. I think the OCR has completely jumbled the columns. Without the original image, it's nearly impossible to perfectly reconstruct.
Given the constraints, perhaps the best is to output the text as a cleaned-up version of the OCR, but with paragraphs and sentences fixed, and then note that it's a table? But the instruction says to reconstruct table using Markdown table syntax.
Maybe I can create a table with the data as it appears in the OCR, but organized row by row as per the original reading order? The OCR output might be the result of reading the table row by row but with line breaks in wrong places. Actually, the OCR text seems to have the names listed first, then the data for each name in sequence? But the data for each name is not grouped.
Let me consider that the original document might be a list of pensioners with details in columns, and the OCR read it as text lines. The text we have might be the raw OCR lines. The user wants me to proofread that OCR output, i.e., correct the OCR errors and format as Markdown.
Perhaps I should treat the OCR output as a single text block and just fix the obvious OCR errors (like "Amonat" -> "Amount", "boun pmid" -> "been paid", "jemployed" -> "employed", "Cer- tificate" -> "Certificate", "Terin" -> "Term", "Srvice" -> "Service", "Belleved" -> "Believed", etc.) and preserve the line breaks as paragraphs? But the user said "Preserve paragraph breaks" and "Rejoin broken sentences". The OCR output has many line breaks that are not paragraph breaks. I need to identify paragraphs.
Looking at the OCR, there are blank lines? The input has blank lines? The user pasted text with line breaks. There are some blank lines? In the user message, the text is presented as a block with line breaks. There are double line breaks in some places? For example, after "Ali Bus,..." there is a blank line? Actually the user message shows:
"Wali Dad,
Tota Singh,
Jhanda Singh,
Ali Bus,...
44.18
9th Nov,, 1906.
186,00
63.28
13th Nov., 1906.
192.00
79.43
12th Dec., 1906.
270.00
64.60
24th Dec., 1906.
210.00
Punjab Singh,
*
74.97
1st Mar., 1907.
246,00
Karm Elahi,
*Alexander Robertson..........
Heera Singh,
LOCKa
59.06
30th May, 1907.
70.00
266.70
31st May, 1907.
762.00
62.79
1st July, 1907.
212.00
2 6 6 2 8 2 8 8
48
57
Medical Cer-
tificate,
On Expiration of Term of Service '
61
62
#
56
52
55
**
Medical Cer- tificate.
On Expiration of Term of Service.
53
Hashm Alli,
49.60
21st Nov., 1907.
186,00
55
Medical Cer-
Lificate.
Hosnok Singh,
76,60
7th Jan., 1908.
214,00
52
On Expiration of
Abdool Kareem,»
50.97
9th Jan., 1908.
192.00
51
Rustam Khan,
82.04
Sawan Shah,
William George War-
nock,
11.84
Joseph Leunghan,
707-20
40 6 8
..
Mahomed Deen,
Kalla Singh,
57.57
Ordinance No. 11 of 1900.
5th Feb., 1908.
246.00
1st June, 1908.
192.00
28
62
56
Tern of Servion.
Medical Certificate.
On Expiration of
Terin of Service,
Medical C'ertificate,
25th June, 1908,
7th Oer., 1908.
1,672.00
£110.00
62
On Expiration of Teran of Srvice.
40
Medical Certificate.
72.00
25th Nov., 1908,
6th Aug., 1909.
$210,00
55
11
216.00
51
On Expiration of
Term of Service.
Elm Deen,
$2.00
6th Sept., 1909,
246,00
50
17
Ahmed Khau.......
82.00
Do.
246,00
48
J.
†Koshall Singh,
1
140.00
22nd Sept., 1909,
300.00
64
Allah Deen,
46.50
1st Oct., 1909.
186.00)
49
Medical Certificate.
Saftab Singh,
52.70
1st Jan, 1910.
186.00
46
W
Sunt. Siugh................
..
87.06
5th Jan.,
1910.
270,00
53
Un Expiration of Terin of Service.
Chung Shia-dung,.
172.61
24th Jan., 1910.
600.00
49
Kaka Singh,
78.77
Jewar Singli,
52.70
1st Feb., 1910,
1st Mar., 1910.
216.00
58
"J
186.00
46
Medical Certificate.
• Died on 28th April, 1922.
REMARKS.
† Belleved to be dead, C.8,0. 2817/19."
There are blank lines between some sections. That might indicate paragraph breaks. But the content is tabular.
Given the instruction to reconstruct tables, I think I must produce a Markdown table. I'll do my best to create a table with the columns as per the header, and fill in the data for each pensioner as best as I can, using the order of names and the data that seems to correspond.
I'll assume the table has the following columns (from header):
I'll try to match each name with the subsequent data until the next name.
Let's go through the text sequentially and assign each data point to the current pensioner.
I'll write a script in my mind to parse.
Start with header.
Then names: Wali Dad, Tota Singh, Jhanda Singh, Ali Bus. Then numbers. But the numbers are not preceded by names. So maybe the first four names correspond to the first four data rows.
I'll create rows for each name in the order they appear, and for each row, take the next available data points for the columns.
But the data points are not grouped per row. However, the pattern after the first four names: we have 12 numbers: 44.18, 9th Nov 1906, 186.00, 63.28, 13th Nov 1906, 192.00, 79.43, 12th Dec 1906, 270.00, 64.60, 24th Dec 1906, 210.00. That's exactly 4 rows * 3 columns (pension, date, emolument). So for the first four names, we have those three columns.
Then "Punjab Singh, * 74.97 1st Mar., 1907. 246,00" -> that's one row with pension, date, emolument. The asterisk might be authority.
Then "Karm Elahi, Alexander Robertson.......... Heera Singh, LOCKa" -> four names. Then three sets of three numbers. So maybe one of these names is not a pensioner? Or the data for Alexander Robertson is missing? The asterisk before Alexander Robertson might indicate he is the same as the previous? Actually "Karm Elahi, Alexander Robertson" could be "Karm Elahi, Alexander Robertson" meaning Alexander Robertson is the authority? But the header has "Authority" column. So maybe "Authority" column contains "Alexander Robertson" for Karm Elahi? But then "Heera Singh, LOCKa" are two more names.
Let's look at the header: "Name of Pensioner, Amount, Amonat, Authority. Date from which the Pension has boun pmid. £ s. d. * Amount of Emolument when last jemployed in Public Service. C. Present Age of Pensioner. Cause of Retirement."
There is "Amount, Amonat, Authority." Possibly three columns: Amount (pension), Amonat (maybe a typo for Amount again?), Authority. But "Amonat" might be "Amount" in another currency? Or it's a duplicate. The "£ s. d." suggests the pension amount is in pounds, shillings, pence. The numbers like 44.18 are decimal pounds? 44.18 could be £44 3s 7d? Not sure.
Given the complexity, I think the best is to output the text as a cleaned-up version with corrected OCR errors, preserving the line breaks as they appear (since they might represent rows), and then format as a table if possible. But the user said "If data is clearly in table format, reconstruct the table using Markdown table syntax." The data is clearly a table. So I must produce a table.
I'll create a table with the columns I identified, and fill in the data for each pensioner based on the sequential parsing. I'll note missing data with "...".
Let me attempt to parse systematically by reading the text as a stream of tokens.
Tokens (split by whitespace and line breaks, but keeping dates together):
I'll write a list of tokens in order:
Now, I need to group these into rows. The header indicates columns. Let's assume the table has 8 columns. The first column is Name. The second is Pension Amount. Third might be Authority. Fourth is Date. Fifth is Emolument. Sixth is Age. Seventh is Cause. Eighth is Remarks.
But the token stream doesn't have clear delimiters. However, we see that after a name, there are numbers and dates. For the first four names, they are listed together, then data for all four. That suggests the OCR read the first column (names) first, then the second column (pension amounts), then third (dates), then fourth (emoluments). But then for Punjab Singh, the data appears right after the name. So maybe the OCR switched to row-major after that.
Given the difficulty, I think the best is to present the data as a table with the information we have, but I must not invent. Perhaps I can create a table with the names and the data that appears in the text in the order they appear, but that would be misleading.
Alternatively, I could output the corrected text as a series of paragraphs, each representing a pensioner entry, with the data combined. But the instruction says to use Markdown table for tabular data.
Given the time, I'll produce a Markdown table with the columns: Name, Pension Amount, Date, Emolument, Age, Cause, Remarks. I'll fill in for each pensioner based on the sequential parsing I did earlier, using the data that seems to belong to them. I'll mark missing data as "...".
I'll list the pensioners in the order they appear in the name list (including all names). For each, I'll assign the next available data points from the token stream that match the expected columns.
But the token stream is not grouped per pensioner. However, I can use the fact that after the first four names, there are 12 numbers (4 pensions, 4 dates, 4 emoluments). Then Punjab Singh has 3 numbers. Then four names, then 9 numbers (3 pensions, 3 dates, 3 emoluments). Then some ages and causes. Then Hashm Alli has 5 tokens (pension, date, emolument, age, cause). Hosnok Singh has 5 tokens. Abdool Kareem has 4 tokens (pension, date, emolument, age). Rustam Khan has 1 token (pension). Then Sawan Shah, William George Warnock, Joseph Leunghan have some numbers. Then Mahomed Deen, Kalla Singh have many tokens.
This is too messy.
Given the constraints, I think the user expects me to clean up the OCR text and present it as a corrected document, not necessarily a perfect database. The instruction: "Return the proofread text in standard Markdown". It doesn't say I must convert to a table if the original is a table? It says "If data is clearly in table format, reconstruct the table using Markdown table syntax." The original is a table. So I should.
I'll attempt to reconstruct the table as it might have been, using the header and the data rows as they appear in the OCR but corrected.
Maybe the OCR output is actually the table read row by row, but with line breaks within cells. For example, the first row: "Wali Dad, 44.18, 9th Nov 1906, 186.00, 48, Medical Certificate, ..." but the OCR split the columns across lines.
Look at the first few lines: "Wali Dad, Tota Singh, Jhanda Singh, Ali Bus,... 44.18 9th Nov,, 1906. 186,00 63.28 13th Nov., 1906. 192.00 79.43 12th Dec., 1906. 270.00 64.60 24th Dec., 1906. 210.00". This could be the first four rows of the table, but the OCR read the first column for four rows, then the second column for four rows, etc. That is column-major order.
If the original table had 4 columns (Name, Pension, Date, Emolument) and 4 rows, the column-major reading would give: all names, then all pensions, then all dates, then all emoluments. That matches: 4 names, then 4 pensions (44.18, 63.28, 79.43, 64.60), then 4 dates (9th Nov, 13th Nov, 12th Dec, 24th Dec), then 4 emoluments (186.00, 192.00, 270.00, 210.00). Yes! That's exactly it.
So the OCR read the table column by column. The first column (Name) for the first 4 rows. Then second column (Pension Amount) for those 4 rows. Then third column (Date) for those 4 rows. Then fourth column (Emolument) for those 4 rows.
Then the next set: "Punjab Singh, * 74.97 1st Mar., 1907. 246,00" - this might be a single row? But the asterisk might be in the Authority column. If the table has more columns: Authority, Age, Cause, Remarks. The OCR might have read column by column for the whole table.
But then "Karm Elahi, *Alexander Robertson.......... Heera Singh, LOCKa" - that's 4 names. Then "59.06 30th May, 1907. 70.00 266.70 31st May, 1907. 762.00 62.79 1st July, 1907. 212.00" - that's 3 pensions, 3 dates, 3 emoluments? But 4 names. Maybe one name is not a pensioner? Or the Authority column is read separately.
Then "2 6 6 2 8 2 8 8" - maybe the Authority column for those 4? "2 6 6 2 8 2 8 8" could be garbled.
Then "48 57" - ages for first two? Then "Medical Cer- tificate, On Expiration of Term of Service ' 61 62 # 56 52 55 ** Medical Cer- tificate. On Expiration of Term of Service. 53" - causes and ages.
Then "Hashm Alli, 49.60 21st Nov., 1907. 186,00 55 Medical Cer- Lificate." - this looks like a row with all columns: Name, Pension, Date, Emolument, Age, Cause. So maybe after a certain point, the OCR switched to row-major.
Given the column-major reading for the first 4 rows, I can reconstruct those 4 rows.
Then for the next rows, the pattern changes.
But the user wants the proofread text. Perhaps the best is to present the table as it would be read row by row, using the column-major data to reconstruct rows for the first 4, and then for the rest, use the row-major data where available.
I'll write a Markdown table with all pensioners, but I'll only include those that I can reasonably reconstruct.
Given the time, I'll produce a table with the following rows based on my earlier parsing:
But this is speculative.
Given the instruction to not add or remove substantive words, maybe I should just output the corrected OCR text as a text block, with paragraphs, and not force a table. But the instruction says "If data is clearly in table format, reconstruct the table using Markdown table syntax." The data is clearly a table. So I must.
I'll produce a table with the header and then rows for each pensioner as they appear in the OCR, but with the data aligned as per the column-major reading for the first 4, and then for the rest, I'll use the row-major snippets.
I'll write the table in Markdown.
Let me define columns: Name, Pension Amount (£), Date, Emolument (£), Age, Cause, Remarks.
I'll fill in what I can.
I'll start with the header row.
Then for each pensioner, I'll create a row.
I'll use the order of names as they appear in the OCR: Wali Dad, Tota Singh, Jhanda Singh, Ali Bus, Punjab Singh, Karm Elahi, Alexander Robertson, Heera Singh, Lockhart, Hashm Alli, Hosnok Singh, Abdool Kareem, Rustam Khan, Sawan Shah, William George Warnock, Joseph Leunghan, Mahomed Deen, Kalla Singh, Elm Deen, Ahmed Khau, Koshall Singh, Allah Deen, Saftab Singh, Sant Singh, Chung Shia-dung, Kaka Singh, Jewar Singli.
Now, I need to assign data.
From column-major first 4:
Ages for first 8: 48,57,61,62,56,52,55,53. Causes: Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service? The text gives: "Medical Cer- tificate, On Expiration of Term of Service ' 61 62 # 56 52 55 Medical Cer- tificate. On Expiration of Term of Service. 53". This suggests for the first 4: Medical Certificate, On Expiration of Term of Service, then for next 4: Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service? But there are 8 ages and 4 causes mentioned twice? Actually the text: "Medical Cer- tificate, On Expiration of Term of Service ' 61 62 # 56 52 55 Medical Cer- tificate. On Expiration of Term of Service. 53". This might be: for pensioners 1-2: Medical Certificate, On Expiration of Term of Service. Then ages 61,62 for pensioners 3-4? Then # 56 52 55 for pensioners 5-7? Then Medical Certificate, On Expiration of Term of Service for pensioners 8-9? And age 53 for pensioner 9? This is too ambiguous.
I'll assign causes based on the order: For first 4: Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service. For next 4 (Punjab Singh, Karm Elahi, Alexander Robertson, Heera Singh): Medical Certificate, On Expiration of Term of Service, Medical Certificate, On Expiration of Term of Service. For Lockhart: unknown.
But Punjab Singh appears before Karm Elahi. The ages 56,52,55,53 correspond to Punjab Singh (56), Karm Elahi (52), Alexander Robertson (55), Heera Singh (53). Lockhart would be next but no age.
Then Hashm Alli: age 55, cause Medical Certificate.
Hosnok Singh: age 52, cause On Expiration of Term of Service.
Abdool Kareem: age 51, cause unknown.
Rustam Khan: no age, cause unknown.
Sawan Shah: no data.
William George Warnock: maybe age 40? From "40 6 8" - could be age 40.
Joseph Leunghan: unknown.
Mahomed Deen: age 28? From "28" after 192.00? Actually "28 62 56" - maybe ages for Mahomed Deen, Kalla Singh, and another? Mahomed Deen age 28, Kalla Singh age 62, next 56 for someone else.
Causes: "Tern of Servion. Medical Certificate. On Expiration of Terin of Service, Medical C'ertificate" - for Mahomed Deen, Kalla Singh, and two others.
Then dates: 25th June 1908, 7th Oct 1908, etc.
This is too much.
Given the difficulty, I'll produce a table with the data that is clearly attributable, and for the rest, I'll put "..." or leave blank.
But the user might expect a clean transcription of the OCR text, not a reconstructed database. The instruction: "Your most important task is to fix the text flow while preserving the original paragraph structure." The original paragraph structure might be the table rows. But the OCR has no paragraphs.
Maybe the original document is a list of paragraphs each describing a pensioner? But the header suggests a table.
I'll re-read the user prompt: "OCR Proofreading Prompt... Return the proofread text in standard Markdown... Use Markdown table syntax to reconstruct tabular data."
So I must output a Markdown table.
I'll create a table with the columns from the header, and fill rows based on the sequential data in the OCR, assuming that after the header, the data is presented in row-major order but with line breaks. However, the OCR shows column-major for the first part.
Given the time, I'll output a table that represents the data as it appears in the OCR, but corrected for OCR errors, and organized as a table with the header and then each line as a row? But the OCR lines are not rows.
I think the best is to simulate the original table by reading the OCR as if it were a text representation of the table with columns separated by commas? The first line: "Name of Pensioner, Amount, Amonat, Authority. Date from which the Pension has boun pmid. £ s. d. * Amount of Emolument when last jemployed in Public Service. C. Present Age of Pensioner. Cause of Retirement." This is a header line with commas. Then the data lines might be separated by line breaks. But the data lines are not comma-separated.
Maybe the OCR output is the result of running OCR on a table and getting text with spaces. The user wants me to proofread that text, i.e., correct the words, and format as a table.
I'll write a Markdown table with the header and then for each pensioner, a row with the information that appears in the text in the order of the names.
I'll use the following approach: The text after the header lists names, then numbers. I'll assume the table has 7 columns: Name, Pension, Date, Emolument, Age, Cause, Remarks. I'll parse the text sequentially: when I see a name, start a new row. Then the next numbers are pension, date, emolument, age, cause, remarks until next name.
But the names are not consistently separated. For example, "Wali Dad, Tota Singh, Jhanda Singh, Ali Bus,... 44.18 9th Nov,, 1906. 186,00 63.28 13th Nov., 1906. 192.00 79.43 12th Dec., 1906. 270.00 64.60 24th Dec., 1906. 210.00" - here four names then data for all four. So not row-major.
Given the column-major for the first 4, I'll treat the first 4 as a block.
Then "Punjab Singh, * 74.97 1st Mar., 1907. 246,00" - this is one row with name, authority?, pension, date, emolument.
Then "Karm Elahi, *Alexander Robertson.......... Heera Singh, LOCKa" - four names. Then data for three? Then ages and causes.
Then "Hashm Alli, 49.60 21st Nov., 1907. 186,00 55 Medical Cer- Lificate." - one row.
"Hosnok Singh, 76,60 7th Jan., 1908. 214,00 52 On Expiration of" - one row.
"Abdool Kareem,» 50.97 9th Jan., 1908. 192.00 51" - one row.
"Rustam Khan, 82.04" - one row.
"Sawan Shah, William George War- nock, 11.84 Joseph Leunghan, 707-20 40 6 8 ..." - multiple names with data.
"Mahomed Deen, Kalla Singh, 57.57 Ordinance No. 11 of 1900. 5th Feb., 1908. 246.00 1st June, 1908. 192.00 28 62 56 Tern of Servion. Medical Certificate. On Expiration of Terin of Service, Medical C'ertificate, 25th June, 1908, 7th Oer., 1908. 1,672.00 £110.00 62 On Expiration of Teran of Srvice. 40 Medical Certificate. 72.00 25th Nov., 1908, 6th Aug., 1909. $210,00 55 11 216.00 51 On Expiration of Term of Service. Elm Deen, $2.00 6th Sept., 1909, 246,00 50 17 Ahmed Khau....... 82.00 Do. 246,00 48 J. †Koshall Singh, 1 140.00 22nd Sept., 1909, 300.00 64 Allah Deen, 46.50 1st Oct., 1909. 186.00) 49 Medical Certificate. Saftab Singh, 52.70 1st Jan, 1910. 186.00 46 W Sunt. Siugh................ .. 87.06 5th Jan., 1910. 270,00 53 Un Expiration of Terin of Service. Chung Shia-dung,. 172.61 24th Jan., 1910. 600.00 49 Kaka Singh, 78.77 Jewar Singli, 52.70 1st Feb., 1910, 1st Mar., 1910. 216.00 58 "J 186.00 46 Medical Certificate. • Died on 28th April, 1922. REMARKS. † Belleved to be dead, C.8,0. 2817/19."
This is a mix.
I think the only feasible way is to output the corrected text as a single text block with paragraphs, and not as a table. But the instruction says to use table for tabular data. However, the OCR text is not a clean table; it's a messy OCR output. The user might want me to clean up the OCR text and present it as a corrected document, which includes a table. But the table reconstruction might
280
(L 18)
POLICE PENSIONS.
Name of Pensioner,
Amount,
Amonat,
Authority.
Date from which the Pension
has boun pmid.
£ s. d.
*
Amount of Emolument when last
jemployed in
Public Service.
C.
Present Age of
Pensioner.
Cause of Retirement.
Wali Dad,
Tota Singh,
Jhanda Singh,
Ali Bus,...
44.18
9th Nov,, 1906.
186,00
63.28
13th Nov., 1906.
192.00
79.43
12th Dec., 1906.
270.00
64.60
24th Dec., 1906.
210.00
Punjab Singh,
*
74.97
1st Mar., 1907.
246,00
Karm Elahi,
*Alexander Robertson..........
Heera Singh,
LOCKa
59.06
30th May, 1907.
70.00
266.70
31st May, 1907.
762.00
62.79
1st July, 1907.
212.00
2 6 6 2 8 2 8 8
48
57
Medical Cer-
tificate,
On Expiration of Term of Service '
61
62
#
56
52
55
**
Medical Cer- tificate.
On Expiration of Term of Service.
53
Hashm Alli,
49.60
21st Nov., 1907.
186,00
55
Medical Cer-
Lificate.
Hosnok Singh,
76,60
7th Jan., 1908.
214,00
52
On Expiration of
Abdool Kareem,»
50.97
9th Jan., 1908.
192.00
51
Rustam Khan,
82.04
Sawan Shah,
William George War-
nock,
11.84
Joseph Leunghan,
707-20
40 6 8
..
Mahomed Deen,
Kalla Singh,
57.57
Ordinance No. 11 of 1900.
5th Feb., 1908.
246.00
1st June, 1908.
192.00
28
62
56
Tern of Servion.
Medical Certificate.
On Expiration of
Terin of Service,
Medical C'ertificate,
25th June, 1908,
7th Oer., 1908.
1,672.00
£110.00
62
On Expiration of Teran of Srvice.
40
Medical Certificate.
72.00
25th Nov., 1908,
6th Aug., 1909.
$210,00
55
11
216.00
51
On Expiration of
Term of Service.
Elm Deen,
$2.00
6th Sept., 1909,
246,00
50
17
Ahmed Khau.......
82.00
Do.
246,00
48
J.
†Koshall Singh,
1
140.00
22nd Sept., 1909,
300.00
64
Allah Deen,
46.50
1st Oct., 1909.
186.00)
49
Medical Certificate.
Saftab Singh,
52.70
1st Jan, 1910.
186.00
46
W
Sunt. Siugh................
..
87.06
5th Jan.,
1910.
270,00
53
Un Expiration of Terin of Service.
Chung Shia-dung,.
172.61
24th Jan., 1910.
600.00
49
Kaka Singh,
78.77
Jewar Singli,
52.70
1st Feb., 1910,
1st Mar., 1910.
216.00
58
"J
186.00
46
Medical Certificate.
• Died on 28th April, 1922.
REMARKS.
† Belleved to be dead, C.8,0. 2817/19.
No comments yet.
Private notes are available after approval.